Facebook

Winnipeg, Canada

Your genealogist descendants will find this and mark your occupation as "koala watcher".

Alan: I thought VAT is for all goods, including digital products. I use FastSpring to sell my software, and they've always charged VAT to my EU customers including the UK. I can't believe the UK would want to change that. https://fastspring.com/tax-management/vat/

Fooled me.

And how did you manage to do that??

Yes, it appears FTB's "Help->Check for Updates" is not recognizing that a new version is available. But downloading directly from https://www.myheritage.com/family-tree-builder and installing will upgrade safely to the new version. The new pedigree view is great, working just like the website. Are there any other improvements in this version you can tell us about?

George - Whereas I like to clear all my hints out and have everything either accepted or rejected.

Yes, Nancy, I do reject them almost all the time as well. It just feels wrong to reject it when the match itself is correct.

I exactly agree with your point. Instant Discoveries should look more like Smart Matches and they'd be more manageable. Most of my Instant Discoveries are also inlaws that I don"t want to add to my tree. With only two options: Accept and Reject, I have to reject them so they won't get added to my tree. They really need 3 options; 1. Accept match and add to tree 2. Accept match but don't add to tree. 3. Reject match.

Love it! Reposting on my time line.

I keep an exact working copy of my sites on my desktop computer, which allows me to test all changes and prototype new sections before I put them up live. I use Scooter Software's Beyond Compare program to sync my sites with my desktop. That has saved me more than once.

I'm curious how you know how many dropouts you have. Are you counting them?

Ann N. Tafel! Love it!

Hi Thomas!!

Harry Yanofsky was the brother of Abe Yanofsky, the chess Grandmaster from Winnipeg (1st grandmaster in the British commonwealth). But I don't see any connection with George's ancestor Morris Yanovsky.

Joel - Sending you an email.

Joel - I'm sure I've already sent you everything I got from Sorin. Feel free to email me.

George Turner - I live in Winnipeg and I'm tracing back all the Mezhirichers that came and formed the Mezhiricher synagogue here in the early 1900s. My grandfather Joseph German was from there. See My Family Research page: www.lkessler.com/myfamily.shtml

Vicki Hoskins We very likely are. Check out My Family Research page, and let me know where you fit in. My email is at the bottom of that page. www.lkessler.com/myfamily.shtml

Did any of them live in Winnipeg?

I've taken a WGS short reads test and a WGS long reads test, and I'm not taking anything else until some company makes a PacBio HiFi test available.

PacBio just published a really good article about phasing with their HiFi reads. https://www.pacb.com/blog/ploidy-haplotypes-and-phasing/

Jo and Ann: May the festival be with you.

You and your wife can beat this, Geoff. We're all hoping for you.

I was planning to go to this event last March, but it was canceled because of Covid.
https://mbchesshof.org/2020/02/23/346/

Some people build bird house for their birds. Maybe you should build koala houses for your koalas.

Jaffa just couldn't seem to figure out the route to mum was the trunk and not that small branch.

Best wishes for you to get better, Vik.

I just discovered that my niece is related to Winston Churchill, Sir Robert Borden and Randy Seaver! https://www.beholdgenealogy.com/blog/?p=3609

So why do you think you got the tongue stuck out at you?

Very colorful food.

Sometimes (fortunately rarely) I do an hour of complex software programming and the power goes out. After a few choice words and a few minutes the power is back on.

I invariably find that doing it over again a second time takes less time than the first and the quality of the result is better.

Not that I want that to happen, but I hope your result the second time was also easier and came out better.

My father was a Canadian farmer.

Do the two younger boys have their eyes on Jewel and Promite?

I think Microsoft 365 is a great deal. For $109 CDN a year, we get the complete Office suite for me, my wife, and my daughters. It includes full versions of Word, Excel, PowerPoint, OneNote, Outlook, Access and Publisher. Plus 1 TB of OneDrive space for each of us.

Not quite as stylish as being on a genealogy cruise, but a close second.

Ha ha, Dave. My analytic abilties are still okay. But if I ever really had good forecasting abilities, I would have made my millions by now and have retired years earlier.

Not late, Linda. Right on time. Thanks.

Oh wow! Thanks, Joel

Someone in elementary school used to call me Louie-Louai because of this song. I forget who that was.

Maybe Thormeo?

If he's becoming a regular, he needs a name.

Discovered Canadian Newspapers got indexed recently and that has resulted in my finding very important obituaries to help fill in my family information.

Tell him about the ladies in your back yard. That might get him interested in making a visit.

I managed to find myself alone in a breakout room twice. Couldn't figure out why because I did shower in the morning. Linda fixed me up though. I like those breakout rooms. It's like at a real conference where you talk to a few random people in the hall between talks.

Gints Klimanis - The technology is less than a year old. No consumer-based company has yet invested in PacBio's Sequel II System. Consumables are still expensive at one to two thousand dollars per test. So maybe in a few years.

That wasn't very nice of them. I've got some mods that I have to make as well. Thanks for mentioning or I might not have noticed for a while.

Oh, and it will need night vision.

Alona - That's a really good idea! Any suggestions for a good model I can place outdoors? Sounds like a good project for next summer (Northern hemisphere)

We can see you. Esther must be lost

I have tried to attract Hopper and Speedy with walnuts. About 30 times last summer, I put 2 walnuts on my deck near my chair in a place I can see from inside the house. They would be gone within 3 days.

I never did catch them in action. Maybe they wake up too early for me.

I did once see one thief, but it was a black raven. So now I'm not sure whether I'm feeding the squirrels or the ravens.

It is my understanding that most companies allow for about 1 mismatch every 100 or so SNPs. That is to allow for misreads and the occasional insertion, deletion and mutation. So you might want to include a column with a 100 length moving average of the number of differences to allow for these.

Which one will get the final rose?

Here's a few of the tips I can give for WATO after my use of it: 1. Make sure you are using Version 2 and not Version 1. The probabilities have changed and are more accurate in V2 for smaller segments under 40 cM. The results between versions can be very different. 2. WATO"s probabilities come from an Ancestry paper and it works best with Ancestry data. Use the standard cM Ancestry gives you after Timber because Timber was designed to give you average cM that match the expected averages for each relationship. 3. Don't worry about endogamy with Ancestry data. Timber does an excellent job removing its effect. But it is a problem with data from other companies. 4. If you are using FTDNA data, recalculate all your matches stripping out all you segments under 7 cM. That will give good numbers comparable to Ancestry and strips out most of the effect of endogamy. There is a tool on the DNA Painter site to help you do this. 5. What WATO can't do for you is handling more than one relationship between two people.

We could use Jennifer as the tie breaker if we knew the email she uses on GEDmatch.

God of thunder thighs.

They wouldn't let their big guy back in if a little one was already there, would they?

I read that the gestation period for kangaroos is about a month, and they stay in the pouch about 6 months. So wait and see if Diamond has a new baby brother or sister 7 months from now.

Leah - I guess it's just a matter of removing those people who have "impossible relationships" like I did with Rob and And in this case. Thanks again. p.s. I've now updated my blog post to reflect your solution.

Leah - Our current theory is that Rob and And's mother or grandmother is related to Gedalia's wife. We're now checking the shared matches on Ancestry between Hypothesis and Rob (and And), and will be comparing them to the shared matches between Hypothesis and the other testers to find testers who only share with Rob or And. Hopefully we can then build up their tree in CeCe Moore style to find the wife of Gedalia. Our situation's really messy, because Moshe has a brother Yehuda who we believe married Gedalia's wife after he died. We have testers under Yehuda who have much more matching between them than they should with Gedalia's descendants. (I won't mention that Yehuda's wife was a sister to Moshe's wife 3. Or that Yehuda's daughter married Moshe's son (from wife 1). - Oops. I mentioned it.) I suppose there is no way to include these multiple relationships in WATO.

Yes, indeed it does!!! I worked on this for many hours and did not see that possibility. Thank you!

Leah LaPerle Larkin This is the Brother tree: https://dnapainter.com/tools/probability/view/945d58ddcd9920f4 This is the Half-Brother tree: https://dnapainter.com/tools/probability/view/be20285652c42756 This is the 1st Cousin tree https://dnapainter.com/tools/probability/view/71025f83676c19e4

Andrew - Yes, that Probability Test Tool is the tool to do the calculation. The difference is I'm not using one tree with a tester that is three hypothesis in the tree. Instead I have 3 slightly different trees where we know where all the testers are in the tree and make one of them the hypotheses. Excellent presentation, by the way.

No Leah - You are not understanding what it is that I am doing differently. I have three different ancestor possibilities, that the two ancestors are brothers, half-brothers or first cousins. I cannot make them hypothesis because they are not DNA testers. Please take a read of my post.

That's why you moved to New Mexico, isn't it?

Thanks Evert-Jan. Didn't know WATO had its own group. Now applying to join.

Ha ha! Is that Diamond or Promite who knocked over all the food getting into the pouch?

Jonny - - See my blog post: https://www.beholdgenealogy.com/blog/?p=3537 Can you think of a way to allow us to use WATO in reverse? i.e. We know how all the descendants share cM. But we don't know exactly how the ancestors are related: Are they brothers, half-brothers or cousins? We would need some mechanism to allow WATO to compare the odds of different relationships of our non-DNA testing ancestors.

Jonny - That's important to know. Then does that imply that it might be better to use Ancestry shared cM with WATO than to use FTDNA shared cM?

Jennifer - Tetiyev, which has no records yet. This is the generation before those who came over, so we mostly only have their fathers' first names from their children's headstones. We actually have 5 possible brothers / half brothers / cousins at the top, all from Tetiyev who took a surname similar to Zaslowsky. The DNA testers are 3 and 4 and 5 generations down, making them no closer than 2nd cousins, 2C1R, 3C and 4C. Some pairs between ancestors, especially between 3C and further do not match each other. I can try WATO as you suggest and put the testers in for one ancestor, and try to add the other as a brother or half and see if his testers fit. But there's are some incompatibilities in doing that. Leah and Jonny based the probabilities on Blaine's shared cM values. Those are an average of all companies. Most of our pairs are FTDNA and which will be higher cM than average. We have some that are Ancestry that will be lower cM than average. Also a third of our pairs are 100% AJ and the rest are a smaller percentage but all have some endogamy which adds on again to Blaine's numbers. This is why Lara's numbers are better for us than Blaine's except that she doesn't give company differences (purpose of my OP) Another complication. We suspect that some of the wives of the 5 may have been related (sisters or cousins) and in one case, A's first marriage may have been with B's widow. And we also have B's daughter marrying C's son and some of our testers are from that line. I don't believe WATO handles multiple ancestral lines yet. But I'll try WATO as you suggest and post what it shows and any hypothesis it may support or refute.

You can't tell from this picture, but from your video of her that you also just posted, she'd be saying. "A little wind can't bother me."

Steve - WATO can only help you determine where an unknown DNA tester is in your tree. It cannot determine if two ancestors of a number of DNA testers are brothers. The ancestors have not DNA tested and we do not know how much they share with all the DNA testers. Thanks for your thoughts on the amount FTDNA adds. You say that's about 60 cM on average. Our initial estimates from the pairs we have that have tested at both companies seem to indicate that Ancestry is about 25 cM lower than Lara's numbers and FTDNA is about 15 cM higher, so we come to about a 40 cM difference. Ancestry's Timber actually increases the difference, because for our data it seems to be taking out on average 55% of the cM for anyone under 90 cM. So a sharing of 80 cM without Timber averages about 35 cM with it, i.e. another 45 cM difference with FTDNA. Because we've got endogamous data, we want to compare with Lara's numbers. Thanks for pointing out Blaine's 3.0 tables. Blaine has an endogamy breakdown, and he has a company breakdown, but his company breakdown is for everyone, not just for endogamous. Blaine's company breakdown does give, e.g. for 2C1R, Ancestry at 112 cM and FTDNA at 136 cM for a difference of 24 cM. But that difference will be higher for endogamous because of the extra small segments FTDNA includes.

Adina - You are correct. This cannot be used to prove anything. But we do want to know which of the relationships is most likely, so that we can come up with plausible theories - just like the WATO tool can do.

Jennifer - We have a number of testers descended from Person A, and a number from Person B. All testers are at FTDNA. We are trying to find out if A and B are brothers, half-brothers, or first cousins. So we can determine pairs of matches that if A and B were brothers, then we could average out the cM of the pairs that would be 2C, 2C1R and 3C and they should be close to Lara's numbers. Similarly, if A and B were half-brothers then the pairs would be H2C, H2C1R and H3C and we could compare those to Lara's numbers. But Lara's numbers are an unknown mix of FTDNA, Ancestry and other companies. So I don't know how much higher our numbers should be because ours are all FTDNA.

Gyorgy - Yes, but unfortunately, neither Lara or Blaine have yet tried to get unweighted values, so there are no shared cM tables to compare the unweighted values with.

Sandy - Actually, I was originally thinking of testers at FTDNA versus testers at Ancestry. But you bring up a good point. There may be an additional difference between a tester at FTDNA and a transfer to FTDNA.

We have always had lots of sparrows as I remember them from when I was very young. They sometimes have very protracted, violent and noisy fights with the squirrels. Haven't been able to get one of those fights on video yet.

We've joined you.

Leandro - Any short read 30x test is fine. I got my long read test from Dante but they don't offer them any more. I don't know who still does.

Unfortunately more coverage with short reads doesn't really improve anything. They still don't cross gaps and are not good at finding most of the insertions and deletions. Better is to combine a short reads test with a long reads test. Even better is to wait for PacBio's HiFi technology of accurate long reads to become commercially available.

It will feel a lot longer this year without the prospect of our usual 2 week getaway somewhere warm.

We woke up to a light covering of snow one day last week, but that melted by noon. We are expecting 3 to 6 cM tomorrow.

Apparently winds across Lake Winnipeg were picking up moisture and causing snow squalls to the East. Berens River, MB got 60 cM over the weekend. And a bit of that must have made it all the way to you.

Just got mine today in Canada. Hooking it under my 2020 B&HF calendar so that it will be ready when the new year arrives.

Surprised that it included a bonus poster of 27 of our favorites.

Still don't know how you tell them apart. Minnie, Boomer and Winston look so similar, even their noses.

Alona - Maybe you should watermark your pictures with a tiny version of your Boomer and her Friends logo in one corner to identify them that they're your beautiful pics, like Judy does with hers.

I would much like the step to create the BAM file be two steps: 1. Assemble Reads de novo 2. Align Assembly to Reference

Pretty good! I had uploaded my combined raw data from five DNA tests. Geneanet put me at DNA Pedigree's tree level 35: R-CTS6, whereas I'm R-BY89987 from my BigY-500 test which is at tree level 40. By comparison, I tested at LivingDNA and they said I am R-Z93 which is only DNA Pedigree's level 30, and my test at 23andMe was even less precise saying I'm R-M417 which is only level 28. FTDNA's Y111 (for my uncle, my father's brother) gave Y-M459 which is just level 27.

Alona - Do you do the same for the roos? What is their idsntifying characteristic?

Two comments:
1. Promite should be signed up for gymnastics.
2. Oy, the stretch marks mom will have.

Michele - I'll bet Alona has a notebook with "cheat" codes.

Hitchhiker's guide to the Galaxy was wrong. We are the 4th most intelligent species on the planet. Squirrels should be ahead of us with mice and dolphins.

Blaine - p.s. Looking forward to your webinar in a half an hour.

Blaine - approximations from EJ's table in his post.

Blaine - I'm agreeing with you in my comment above, with mathematical reasoning.

Scott Wilds - The child has one allele that is the from the parent and is the same. That will have the identical contribution to making a match by chance. The other allele of the parent and child are independent of each other.

Blaine - It does help for the larger segments. e.g. If a 12 cM match has 10% chance of being false, then both parent and child have 10% * 10% = 1% chance of both being false, which is significantly less likely. But you're right about the smaller ones. e.g. If a 5 cM match has 90% chance of being false, then both parent and child have 90% * 90% = 81% chance of both being false, which is still very likely.

I'm surprised by this table, because I would think that the probability of being IBD would be completely dependent on the number of SNPs. That's because the probability of matching by chance is computed from the average probability of heterozygous SNPs on the person and the person's match combined with the probability that one of the alleles on each SNP of one person randomly matches an allele of the other person. Since neither of those probabilities is dependent on cM, I see no reason why cM should have anything to do with it.

So will you call her Mumma Bear or Minnie now?

Did Ned eat the full container?! You'll soon have to buy roo food in 100 kg bags.

Ann Turner recently got her AncestryHealth raw data results. They use a different chip and maybe that raw data format is different.

I'm 100% as well. Same range as Sara: 98-100%

I just got mine today. They increased my Ashkenazi Jewish from 92% to 96%. Central Europe went down from 7% to 3%; I lost my <2% Rest of Africa, and I gained a <1% Irish. So my Jewish portion is now 100% at Ancestry, 99.2% at 23andMe, 96% at Family Tree DNA, and just 83.8% at MyHeritage.

Still none. 😢

Missed it the fist time. But got it ordered this time.

Paula Berman - If in the future, I discover any connection I have to Olga, I'll be sure to contact you.

Altona, Manitoba is a town 100 km SW of where I live. https://en.m.wikipedia.org/wiki/Altona,_Manitoba

I believe I found your ggrandmother on Geni. It gives her first name as Olga rather than Golda. https://www.geni.com/people/Olga-Berman/6000000030664329291

Paula - I have German ancestors who were possibly Herman who we thought were from Mezhirichi. Vik Chymshyt obtained some records from there but we didn't find any German/Hermans in them. A couple of months ago, I found some German families in Tuchyn on Jewishgen under the Ukraine Revision lists from the 1850s. Tuchyn is only 10 miles away from Mezhirichi, so maybe my family or their relatives were originally in Tuchyn. And maybe my German/Herman's are related to yours. Please see My Family Research page and check out my Herman section. If anything there seems familiar to you, please let me know. My email address is at the bottom of the page. www.lkessler.com/myfamily.shtml

I guess the koalas are not accepting food from you like the roos are. They must have plenty of eucalyptus leaves to eat. Any way for you to get them to come down off the trees for you? What if you collected a big bowl of eucalyptus leaves for them?

Evert-Jan Blom - A neighborhood in my city is named Saint Boniface https://en.m.wikipedia.org/wiki/Saint_Boniface,_Winnipeg

No. Number 8. BDI and Sargent Sundae are both equally valuable.

18. Number 8 should count as 2. I've taken a trip to Grand Forks but haven't took one.

Blaine - I'm not meaning to hijack this thread, but you should try Kevin Borland 's almost-ready-for-prime-time Phasing Tool at Borland Genetics for your phasing project.

Found his obituary.

Yes, Strewchuk's was at 1733 Main St. and was called North Main Groceteria. Next door at 1731 Main was Thomson's Grocery. My great-uncle Harry Zelickson used to own 1731 Main in the 1950's when it was named Hi Way Grocery. In the 60's he moved to 1375 Main between Cathedral and Bannerman which he called Zelickson's Grocery.

Looked in a 1965 Henderson's Directory. It was his last name: William Jurek. 412 Hartford at Andrews. Store was called Hartford Grocery.

Flintstones, Yogi Bear, Bugs Bunny, Supercar, Jonny Quest. Wash and Ash reading the Saturday morning comics on the radio. Sunday night was always The Wonderful World of Disney followed by The Ed Sullivan Show.

Hartford and Andrews, southwest corner. We knew it as Jureck's (not sure of the spelling but I think that was the owner's first name). The store was there until about 1965 or 1970. We used to go there after elementary school. Mom gave me pennies to buy candy.

<td colspan="2" align="center">

Wynn Edel G That's the one! Love the car parked in front. I recall they had an area of pool tables inside. And now I see the sign says "Billiard Lounge"

Did you go on the walk?

Troy - That's a nice old shot of the mall. But if the bowling alley was there then, it would be directly behind the Loblaws. The Loblaws was a separate building in the southwest corner of the mall.

Alternatively, or in addition, it would be nice if they could give us the post-timber longest segment.

70 pounds in only 8 months is a tremendous achievement!

Victory, EP, WKCI 1974

You were long suffering until 2016.

Myheritage has a contest on right now for enhancement and colorization that you might want to enter: https://blog.myheritage.com/2020/08/show-us-your-enhancedandcolorized-photos-for-the-chance-to-win/

David Huen - Still, 94% is pretty impressive.

Either you are 2.5 meters tall, or Vegemite is a lot smaller than I imagined. Are Charlie and the other boys that much bigger than all the girls?

You beat me to Ancestry by 51 days. But I've been researching longer. 🙂 I could use the U.S. Discovery membership. Currently I am only subscribed with a Canada Deluxe Membership.

That's what makes this filter valuable to us.

Jennifer - Nothing wrong with them building their trees using a reliable source. And were any of those people new cousins you did not already know or have in your tree?

Jennifer - Yes you could search trees at Ancestry one by one yourself, looking manually for common ancestors. But you don't have a pre-selection of Jewish trees to do this with. Our DNA matches are mostly Jewish. And the Common Ancestors DNA filter does this automatically for us.

My question is how many total matches do you have when you apply just the Common Ancestors DNA filter. I'm guessing you'll have a small number like me (18) rather than a larger number like Blaine (431 prior to purge, 362 after).

Blaine - In your (excellent) article, you said: "I have 69 matches in the range of 6-7 cM that have a Common Ancestor designation, out of 431 total (so 16%)."

By comparison, even though I've got so many more DNA matches than you, I only have 18 matches with a Common Ancestor designation (I'm curious how many Jennifer has). This is likely because my documents in Romania and Ukraine only go back to the 1800's and I can't build my tree more than 5 generation back. That's why I see the Common Ancestor filter as very valuable.

My lowest match of those 18 is a 3C1R who shares 18 cM on 2 segments. So I have a good hint at 18 cM and nothing useful below that.

https://thegeneticgenealogist.com/2020/07/17/losing-distant-matches-at-ancestrydna/

Blaine, The smaller DNA matches for both Jennifer and myself are mostly people with significant European Jewish ethnicity.

Ancestry's Common Ancestors DNA filter will only compare trees of people who are DNA matches. So all these hundreds of thousands of Jewish testers' trees are compared to ours. Through that filter, I found 3 relatives who I was able to determine my connection with, and I was able to add them to my tree and also a connections, not DNA-based connections.

I in no way will ever attempt to say that the small amounts of DNA we share prove the connection. I know that's absolute hokum. The matches may be false or come through different common ancestors than the one the Common Ancestor filter identified.

If Ancestry had a Common Ancestors filter that would compare my tree with all other Jewish trees at ancestry, then I'd say get rid of all my matches below 40 cM. But since these matches are run through that filter, they remain valuable to me, because every so often, the Common Ancestors filter will find another one.

https://www.beholdgenealogy.com/blog/?p=3515

Those of us with endogamy didn't lose as much as those without. I had 194,627 matches before the purge and have 139,471 after. So I lost 55,156 or just 28%,

I did not realize the SNP prefixes indicate who discovered them: "SNP Name Prefixes: The letter prefixes of the SNP label typically represent the laboratory or entity which identified or published the SNP marker. BY and FT => Family Tree DNA laboratory L => Thomas Krahn at FTDNA, primarily from his Walk On the Y project there A => Thomas Krahn at his YSeq laboratory Y => YFull Team (Russia) YFS and YFC are SNP candidate names YFull uses as preliminary designation for new mutations under study FGC and FGCLR => Full Genomes Corp Z & DF => ISOGG or wider genetic genealogy community M => Stanford University and other academic researchers MF => 23MoFang ( 二十三魔方生物科技有限公司 ) - a Chinese genetic genealogy testing lab company CTS => Chris Tyler Smith who cataloged many Y mutations using 1000 Genomes Project data. For a more complete list of SNP prefixes, see the bottom of the ISOGG Tree page (isogg.org/tree) The numbers following the prefix are typically just a discovery sequence number and have no biological meaning."

If today all you did is hold yourself together, scan 40 documents, catalog 100 photos, remove 30 items off you desk, clean out 100 emails from your inbox, and sat outside for 15 minutes, I'm proud of you.

Thanks for that link. I hadn't heard of Genetic Homeland before.

Blaine - You said "extreme endogamy". See: https://m.facebook.com/story.php?story_fbid=10157700498294639&id=572834638

Does she come up to you with Promite in her pouch? If so, that's way more than just trusting you.

Endogamouse

Supercalifragilisticendogamistalodocious

Today we can officially welcome Jay to the family. Wishing everyone all the best.

CIBC - Could you please add Ukraine as well.

I have 6 second cousins at Ancestry whose largest are 58, 38, 34, 25, 23 and 13. The total I share with them (in the same order) are 202, 96, 73, 162, 159, and 99. What I find most interesting about this is that Blaine's Shared cM 4.0 at dnapainter says a 2C should range from 41 to 592 cM with a mean of 229 cM. All my values are below the mean. I have 100% endogamy. What this seems to show is that Ancestry on average is over-downrating my shared DNA with their algorithms.

"Independent research from genetic genealogists suggests that 15 cM is the threshold where segments can be assumed to be IBD with reasonable confidence"
https://cruwys.blogspot.com/2016/01/autosomal-dna-triangulation-part-2.html

"Half-identical matching segments of 15 cMs are mostly IBD"
https://isogg.org/wiki/Identical_by_descent

Debbie - Compare your raw data over a stretch where you match someone, and over another stretch where you don't. You'll likely see about 99 out of every 100 alleles half-match in the first case, and 95 out of 100 in the second. There is quite a distinct separation between the two, which is why matching works so well right now.

Adding extra markers to compare may increase the SNP density, which will help for smaller segments, but segment matches of 15 cM or more are already almost all IBD matches with current microarray technology and WGS can't improve on that.

Debbie - Blaine has in the past referred me to this paper that says WGS will not provide much improvement for relative matching:

Relationship Estimation from Whole-Genome Sequence Data, 2014, Li, Glusman, et al.
https://journals.plos.org/plosgenetics/article?id=10.1371%2Fjournal.pgen.1004144
Gives only a 5% to 15% improvement in WGS over microarray for matching purposes

So true.

The one really good thing about the year is the onset of virtual conferencing. I've been to several virtual genealogical conferences already this year that I wasn't able to attend when they were physical. Also I've had several Zoom meetings with family and friends, some of whom I hadn't seen in years.

I hope after everything gets normal again, that these virtual additions to our lives will continue.

That is way more than impressive!

I sometimes wonder why genealogy doesn't really promote recording things like handedness, eye color, hair color, wore glasses, height, weight, blood type, allergies and other physical attributes. Those would be interesting to track through the generations and to figure out who passed down what.

Aaron - I completely understand. It's best to handle 95% of the cases efficiently than to bog everything down trying to include some of the exceptions.

Aaron - Yes, I was thinking of changing the ones with the issues, but after using your GEDCOM search and not finding them anyways, I saw it wouldn't have helped. Your search could easily strip any parenthesis or quotes from around names and split up names containing a slash. The real problem in my case are that my forenames and surnames are translated from Romanian or Russian. And the documents I found even spell them differently, so I have to pick just one variation for the name. Someone may have the same ancestor, but chose a different spelling for the name. When I'm looking for my names in a database, I look at anything the sounds the same. Daitch Mokotoff soundex is good for this. And a first name like Joseph can not only be Joe, but has ethnic variations such as Josub, Iosub and Yosef. Also, I see you must be checking other facts to verify the name e.g. Joseph German has a few entries when I search, but they are different people so you correctly don't include them. That is a problem because birthyears from the Eastern European documents often have to be estimated from ages which are as poor as ages on our North American Census documents and often give a several year range. And if you use place names, they are also inconsistent. I've seen a hundred different ways to spell Mezhirichi. Checking only ancestors will work only if two cousins have researched to the same ancestral generation required. If one has fallen one or two generations short, a match won't be found, even if the other has the descendancy correct down those one or two generations. So there's just a few of my thoughts on why some people like me will not have any matches.

I completely understand why I have none. This is the best I can get my pedigree using all the records available, and some of my 4th and 5th generation ancestors hadn't adopted surnames yet. When I manually look up each of my ancestors with surnames in your "Search all GEDCOMs" tool, I come up only with my own.

Thanks for the explanation Aaron. So you're just comparing pedigrees to pedigrees then, which makes sense when you think about it.

Koalas have 16. 8 pair.

Kangaroos have just 20 chromosomes. 9 pair of autosomal and the sex pair.

I have no matches with a 4000 person tree. But my Ancestry only goes back 5 generations due to lack of records in Romania and Ukraine. Also I have few close matches at GEDmatch that would be within 5 generations. The ones that are, either don't have trees uploaded or don't have trees that connect to me.

I used one of those at work back then. 30286 processor with 20 MB hard drive. DOS. Floppy 5 1/4" disks. Wordperfect. Lotus 123.

Gerard: Those look like very small file sizes for a 30x WGS, which should have a BAM and Fastq each about 100 to 150 GB. Are the file sizes shown incorrect, or are those not full BAM and Fastq files?

Eugene - If your matches are direct comparisons from your match list, then it seems like there's a problem with MyHeritage's shared matches.

Eugene: Those are quite long segments not to match either parent. Does either parent match any of those people on those segments, but smaller than 15 cM?

"Top 3". In terms of what?

Really! What other wildlife habitate urban Toronto backyards? In Winnipeg, our daily lives interact with squirrels, rabbits and ducks.

I know Hawaii has wild chickens. And my friend in Australia has regular kangaroo and koala visitors.

Now how are you going to get rid of the raccoons?

"Snowdon is the highest mountain in Wales, at an elevation of 1,085 metres above sea level, and the highest point in the British Isles outside the Scottish Highlands. It is located in Snowdonia National Park in Gwynedd. It is the busiest mountain in the United Kingdom and the third most visited attraction in Wales; in 2018 it was visited by 558,000 walkers, with an additional 140,000 people taking the train. It is designated as a national nature reserve for its rare flora and fauna." "In winter, Snowdon often has a covering of snow (giving rise to its English name) ...The slopes of Snowdon have one of the wettest climates in Great Britain, receiving an annual average of more than 200 inches (5,100 mm) of precipitation." - So you're lucky you got a beautiful day.

Eugene: Excellent work. The key result out of your analysis is that only 4 segments out of 66 = 6% don't match in the parent and are likely false matches in the child. But it gives an indication that individual segment matches are not that bad at MyHeritage. You seem to have a category IIa and IIb. The IIa is clear enough, but the IIb seems to indicate that you think you have matches to some of these 20 relatives through both parents. Why not check the other parent to see if you do. I would expect it would be a rare occurrence at the segment level. Category III are all likely valid. What is happening here is that at the end of segment the child got, they are randomly matching through both chromosomes to both chromosomes of their match, where their parent is not. So the child has a bit of random matching at one or both ends. Remember that a by chance match can be up to 15 cM, so you can have up to 15 cM of random match at either end of a real match. You say the largest extra that you have is just 2.2 cM, and that much extra in no way disproves that these are real matches passed down from the parent. Now take a look at, say 10 of the people who your nephew matches but he doesn't match either parent. Look at those matches in a chromosome browser and see what the distribution of segment lengths are. There shouldn't be any larger than 15 cM. And you'll likely find at the 8 cM threshold, indicating they may be false or that their parent's match might be below the threshold for includance. Actually, your results here (4 out of 66) are better than I would have expected from MyHeritage due to their imputation and splicing.

Here's something else you can do. Find someone on your nephew's father's side that your nephew also matches. Then go to your chromosome browser and include all 3 of them. See how many individual segments your nephew matches that his father doesn't. Do this for about 10 people on his father's side and 10 on his mother's side. That will give you an idea of how good MyHeritage's individual segment matching is. I'm assuming that you/your nephew are not from an endogamous population. If you are then you could have people you match through both parents, but neither parent alone has enough to match. But you are correct. You cannot ever assume that people who don't match one parent, must match through the other. A false match is always a possibility. In your case, with one parent tested, you can use shared matches and/or MyHeritage's clustering algorithm to find groups of matches who don't match the parent tested.

So true. My wife and I were following a U-Haul truck today that had markups all over it advertising its Winnipeg office. But it had Arizona plates. We couldn't figure that out.

Linda - Yes, that's a possible reason for some people. In my case, I have known relatives who DNA match me on all 4 of my grandparents' sides. But none have trees at GEDmatch. Having Eastern Europe roots that only have records to the mid 1800's limits my tree to 5 generations, and that's the problem in my case.

Be happy that you have matches. I have a tree of 4000 people and zero GEDmatch tree matches.

They could be from the pseudoautosomal region which are the tips of the X which are the same as the tips of the Y and can sometimes recombine with them in males. Because they are they same, it might be arbitrary as to whether the variant was assigned to X or Y. https://en.m.wikipedia.org/wiki/Pseudoautosomal_region

Congrats! Is this your 4th grandkid now?

It is encouraging that there are big investors who are willing to pay that much for Ancestry. They obviously see new potentials for growth and will push new innovations now that DNA testing seems to be approaching saturation. Maybe the new NGS health offering is one of the new company's ideas. I'm looking forward to seeing what they come up with in the next few years.

This article has more info. They are being very careful not to give any details about their "new innovation in advanced genetic sequencing automation". I doubt the user will ever get to see the raw data from the test. "Quest Diagnostics Launches Automated Next Generation Sequencing (NGS) Engine To Power AncestryHealth®" https://www.prnewswire.com/news-releases/quest-diagnostics-launches-automated-next-generation-sequencing-ngs-engine-to-power-ancestryhealth-301104654.html

Time to collect some DNA?

Thinking along these lines, what if we aligned the 2 sets of fastq files together using a single pipeline? If we could somehow identify the test each read was from, we could learn which test maps better and compare error rates, variant and indel identification and more. Has a combined assembly like this ever been done and analyzed before?

My uncle and I have a GD 2 at Y111.

Adelaide, Australia (hi, Alona) is 15,263 km = 9,484 miles from Winnipeg as the Crow flies. Same cruise that took you to Perth, Judy.

Jan - So now there's two of us. Cyndi should add "File by repository" to her survey. It saves a lot of time not having to separate and reorganize items from a collection.

Cyndi: Effectively, yes.

None of your "File by" choices. I file by who or where I got the record from.

This got me wondering if the smaller matches are good hints for me, in an endogamous population.

Ancestry classifies me as 100% European Jewish. I can compare my ethnicity with each of my matches and they tell me what % European Jewish my match is. I did this for my closest 20 matches, and my first 20 matches at 40 cM, at 20 cM, 15 cM, 10 cM, 9 cM, 8 cM, 7 cM and 6 cM. I marked down what percentage of European Jewish they had. Then I sorted each group of 20 highest to lowest. I get the table below.

Of the 180 matches I checked 179 had some European Jewish Ancestry. Over half of the matches also had 100% European Jewish ethnicity and many of them have 50% or more.

There is a much greater chance that I might find a connection to someone with European Jewish ancestry than someone without any, so these are good hints. Using ancestral surname and place filtering tools, I might find that some of these people are relatives and they can help me extend my family tree.

Does that mean that we share DNA? Not necessarily. The matches, especially the small ones, may be false matches.,

Or we may actually share DNA. but the segments we share may not be coming from the common ancestor we found, but may be from another more distant line that we’ll never find, or it may be (especially in my case) general background noise from distant ancestors due to endogamy. We don’t know and cannot tell.

None-the-less, these matches are hints that might connect you to a relative.

Kalani - Okay, so yours is 12 generations back related 4 ways, and mine is 10 generations back related 3 ways. Doesn't make much of a difference when neither of us have records going back more than 5 generations.

Kalani - Beat ya by 0.7! I know you have as much hope of figuring out yours as I do mine.

Mine is 104.3 cM on 12 segments, largest 14.5 cM. He turns out to be my 19th closest match out of 15,580 matches at MyHeritage. After my uncle and my 1C1R, my next highest match, #3 is 141.4 cM. My #3 match is said by MyHeritage to be a 1C2R to a 2C1R. My new #19 match is said by MyHeritage to be a 2C to 3C1R. I don't know how any of them from #3 to #19 and beyond are related to me, meaning they cannot be as close as a 3C. MyHeritage does not compensate enough for endogamy. They are all likely 4C or further. Other than my uncle and my 1C1R, the only other person at MyHeritage out of my 15,580 matches that I know my actual relation to is a 2C2R who shares 79.4 cM with me. Isn't endogamy fun?

David Pike has a good writeup and references along with his ROH tool. https://www.math.mun.ca/~dapike/FF23utils/roh.php

How can you tell?

I think they'd do well for Canada in a squirrel Olympics.

Both my parents were born here, but all four of my grandparents were born in Europe, so I'm not sure how to classify them. Would 1900 - 1910 be considered recent?

Laurel Wilson - and I bet most people who DNA tested at Ancestry will have a 6 or 7 cM match with them.

I wouldn't doubt that a full tree can be developed with online research for the (fake) Vanderschmidt and Fernsby families.

Rosemary - There is nothing wrong with researching one of your DNA matches and through documentary records, finding out that they are related to you and making connections to them. That, after all, is why we as genealogists DNA test in the first place. We are trying to find new cousins that we can connect our tree to who may give us added information about our own ancestors.

What is wrong and what is confirmation bias is then assuming that the small DNA match was IBD and was passed down to both of you from the common ancestor that you found.

The small match could very well have been passed down to you from a completely different common ancestor. Or it might have been a false match (not IBD at all), and you actually don't share any DNA at all with that person.

As long that DNA small match assumption is not made, with the claim that the DNA match proves your connection through that common ancestor, then you're fine.

Will Winston's charms finally attract Boomer? Or does Boomer have her interests in another? Come back tomorrow for another Days Of Our Koala's Lives.

Tamura - The 1000 is likely an underestimate. I heard that we have more Canada geese in my city than people. Here's a picture I took today not too far from my house looking in just one direction, and there were many more and multiple occurrences of this on my ride on Tuesday.

David Severman - I love your Avatar.

Very interesting thinking. I'm on GEDmatch and I have an account at MyHeritage and use the same email for both. But I didn't get one of the phishing emails.

You may be right that only people who uploaded to GEDmatch from MyHeritage might have got the phish email. My uploads to GEDmatch were from Family Tree DNA and 23andMe.

I really love the hostas in my front yard. They are all on the north side. And our climate is colder than yours and they are doing wonderful. Never have to water them. Here are my Sum N Substance Hosta with my False Spirea. You can also try Purple Emporer Sedum which I have there as well but have been overgrown by the other two. What I really love about the Spirea and Sedum is that their flowers really attract the bees which we all need to help along. Very little maintenance to all this. Just an hour or two of trimming and cleaning up every Spring.

Thanks for letting us know this is coming. I'm looking forward to it, but no, I don't have it yet.

It's a Canadian classic.

Counting 1000 geese is easier than counting 100 butterflies.

Alona - I personally like a 12 month calendar. You could just include koalas and roos that weren't in last year's calendar, and have Boomer on the cover each year. Then we'd have to keep our calendars each year so as to collect them all. And you'll have to keep taking great pictures of all your new family that visit so you can keep this going for years.

Thinking logically, the cover is only for marketing because it never gets looked at. The 12 months each get to to be viewed each for a month. So you need the cutest, most adorable and irresistible picture you can find to make your calendar a must-have. I think you've got some better ones, like a mama and a baby. And if the cover is that good, then it might prompt leaving the cover page up on the wall for a few months into the next year. If this calendar is only Boomer, then the cover has to have him doing something cute, or with another koala.

Anne: Info is here: https://fgs.org/conferences/registration/

You'll find out soon.

Thanks Thomas & Diane. I guess I never made it to the Tier 1 of speakers who get their travel expenses paid for.

Gary - We're lucky we had great promoters in WInnipeg and good venues for them.

I participated in a Facebook meme 3 years ago, listing 9 artists I saw in concert, and 1 that I didn't, and Carole King was my "didn't": https://www.facebook.com/beholdgenealogy/posts/536105356779257

Diane - But you don't have the expense of travel to the site, hotel, transport and food. So your overhead for that gig is much less.

Here's an article I wrote a year ago about small segment matches with Blaine on GEDmatch. https://www.beholdgenealogy.com/blog/?p=2939

We were in Sydney, Australia in 2013 for 4 days, followed by a 7 day cruise (our first cruise ever), followed by 4 more days in Sydney. We found out Carole King was performing in Sydney while we were there, but when we were on our cruise. We almost canceled our cruise to see her (but we didn't).

Saw it a day too late. :-(

Glad it finally rained.

Everybody already has your email address. You get spam, right? Now you'll get a bit more.

What Yvette said, and also I'm not sure what a hacker would do with an IP address, other than find your general location, especially when your actual name and address are in the database. Most people get VPNs because they don't want a website they are visiting to know about that visit.

Clive Moon - Most sites today that know what they are doing use one-way hashed passwords. So the password will not be available to hackers. It would be nice, actually important, for GEDmatch to assure us that they store our passwords securely and not as text.

It now says "maintenance"

Paddy, thanks for explaining. Your cM table has the same ranges on each line as the Mb table where 0.8 Mb = 1 cM. Okay. That's fine then.

But I still don't think generally the distant ancestors will move down the list and the recent ones will move up. I think it will be exactly random as to which move up or down.

Take any size segment, e.g:
8 Mb on average = 10 cM
One extreme: 8 Mb = 100 cM
The other extreme: 8 Mb = 1 cM

So with one extreme the segment will move up your list, and the other down your list. But on average, each segment will stay in the same place.

Here's looking at you, kid.

Paddy, Sorry, I don't get the reasoning behind your statement "distant ancestors will generally appear shorter on the centiMorgan scale than on the megabase scale; segments which descend from more recent ancestors will generally appear longer on the centiMorgan scale than on the megabase scale."

To me, distant ancestors generally have shorter matches. Closer ancestors have longer matches. Whether cM or Mb is longer and shorter just depends on where the segment is on the genome.

And it would make for easier comparison if the ranges in your cM and Mb tables are the same.

Debbie - those names reflect that they are likely people's private research kits.

The PacBio HiFi SMRT reads are getting there. They're accurate enough. They just have to be a bit longer. https://www.pacb.com/smrt-science/smrt-sequencing/smrt-sequencing-modes/

Ted, sorry. I don't know.

Debbie - I don't think I ever mentioned a "View As". I've got what you've got: the m_Louis_Kessler.csv file with the same fields you mentioned. I simply analyzed and made a histogram of the "Shared cM" field.

Randy, Blaine: As a programmer, I've always known this technique as "scraping", as in "scrape" the data off the website. So it should be a "scraped match list" made by a scraper. The word scrap/scrapped means to throw away. Isn't English odd though: scrape -> scraping -> scraped -> scraper scrap -> scrapping -> scrapped -> scrapper

Blaine - Yes, but I can imagine that the average person without endogamy might have a shape more like Shelley's, with fewer larger matches, and many more false smaller matches. What we could be seeing here is the effect of Ancestry's Timber algorithm on endogamy. If they didn't trim my matches and my distribution had the same shape as Shelley's, I might have hundreds of thousands of matches in my smaller groups.

And just for completeness, this is my histogram of how Ancestry rounds them:

Shelley - Hmm. That's interesting. Yes mine is a bit flatter than yours. Here's two histograms with 0.5 and 1.0 buckets.. Yes, maybe it's flatter due to my endogamy.

Maybe this will help. From a DNAGedcom download I did previously, I can see that 6.4999 is displayed as 6 and 6.5 is displayed as 7. The smallest value they include is 6.0. Here are my statistics for my 192,306 matches:

I had to document my feelings about all this. https://www.beholdgenealogy.com/blog/?p=3515

There are also 1.8K members in the Chromosome Mapping of Ancient Bloodlines group. It's horrifying! I wonder if they're the same people?

Blaine - But it's what we mathematicians do. Squabble! It doesn't have to be meaningful.

Blaine - I'm just math-ing it Blaine. I believe most small segments under 7 cM will be 10 to 25 generations old. I don't believe that most of them will be over 20 generations old and I really don't believe that a 10 cM segment will be 32 to 52 generations old.

Genealogically, it doesn't make a difference to me either way, since my Romanian and Ukrainian roots only allow me to research 5 generations max, and I can't bias my maths down that far.

Debbie - Yes, which is exactly why I can't believe there isn't a shorter path of a 10 cM segment to the common ancestor than 32 to 52 generations!!!

(See my reply to your comment below for more discussion on this)

Debbie - I have read Ralph and Coop many times. But I've never been able to reconcile their statement you quote above that the average age of a 10 cM shared segment is between 32 and 52 generations.

The probability of a 10 cM segment recombining in 1 generation is: 9.05%. The probability that it recombines in 20 generations is (1-0.0905)^20 = 85%. The distribution has a mean of 10.5 generations.

If it recombines, then that 10 cM segment no longer exists. It has died. It is a dead segment (like a dead parrot). It can no longer match anyone with that same 10 cM segment.

So R&C are saying that a 10 cM segment survives 32 to 52 generations down one descendant line and the same down a second descendant line to be the AVERAGE number of generations needed to match.

I just find that to be an unbelievably high number of generations and I can't stop from thinking that there has to be a closer MRCA for that 10 cM segment along some other ancestral path.

Paddy - Yes exactly. It is the specific ancestral path that we are dealing with. And the simple math of recombination (i.e. a 10 cM segment will recombine almost 10% of the time) states that the majority of segments will be under 20 years old.

CeCe Moore is speaking on Sept 2nd as part of the Virtual FGS Conference. Her topic will be Strategies of "The Genetic Detective" in which I'm hoping she'll give an insight into the tools she uses. https://fgs.org/conference/2020-conference/

Nicole Dibben-Gamelin - No, BeenVerified seems to be USA only. I've looked around for a good Canadian equivalent, but haven't found one. Seems you can find out anything you want about Americans, but we Canadians don't exist.

Speed and Balding's paper does say they are using the Decode Genetic Map which specifies male and female recombination rates over 2,667 Mb across the 22 autosomes. So even though they are giving results in megabase pairs, they are using correct recombination rates over each Mb for the simulation.

Since 1 Mb is approximately 1 cM, if you use the cM value as a lookup, it should give a reasonable estimate, e.g., you may have a 5 cM segment which is only 1 Mb long, but as long as you lookup 5 Mb in their table and not 1 Mb, you should be okay.

I see among your revisions, you bring up Speed and Balding. I am one who believes Speed and Balding's generation estimates are too high. You might be interested in seeing my 3 different mathematical analyses of segment age that I compare with Speed and Balding in my articles:

Revisiting Speed and Balding:
https://www.beholdgenealogy.com/blog/?p=2338

Another Estimate of Speed and Balding Figure 2B:
https://www.beholdgenealogy.com/blog/?p=2389

The Life and Death of a DNA Segment:
https://www.beholdgenealogy.com/blog/?p=3057

Kevin Borland visits Speed and Balding: https://www.beholdgenealogy.com/blog/?p=3420

I should mention that both Debbie Kennett and Andrew Millard disagree with my analysis and conclusions and we have agreed to disagree.

Even so, I still agree with Blaine that small segments are "poison" because (1) they are often not IBD matches, and (2) they still can be 10 or more generations back. Once you get that far back, it could come from any ancestral path, not just the one you are theorizing. The only way to verify is a Jim-Bartlett-like tracing the triangulation group back generation by generation using DNA matches at each cousin level all the way back.

Lot of good hockey players in the Koivu family.

Maria - You can find every single ThruLines that Ancestry finds, plus every other one that they don't simply by entering each of your people in your Ancestry Family Tree into Ancestry -> Search -> Public Member Trees. Never say "never".

Sorry Ashley, I don't mean to come off strong. But really, if you are using your 6 cM and 7 cM matches (or even up to 20 cM) to try to find connections on your family tree, it will be long and mostly unrewarding task for you. Here's my alternative suggestion, that can do the same task and will likely have a much higher success ratio. Go to your Ancestry Tree. Get your list of all people in your tree. Go to Ancestry -> Search -> Public Member Trees. Enter each person in your tree, one by one, into this search and find all the Ancestry Member trees that contain your relatives. This is much better than trying to use small DNA matches to find possible relatives who might have trees that connect to yours.

Ashley Bens - Yes, that is my intent. I have 192,000 matches. I'd sooner spend my time investigating 1,920 possibly identifiable matches than 190,000 poison matches (to use Blaine's term). In fact, I don't mind, and even like 23andMe's limiting to your closest 2000 matches.

To be honest I always thought Ancestry's low limit of 6 cM was ridiculously low. 15 or 20 cM is more reasonable for the general population. For those with endogamy like me, there's about 40 cM of background noise added to each of my matches so I don't usually look at anyone under 50 cM.

I have 192,306 matches (a large number due to Ashkenazi endogamy). 54,498 of those are 6 or 7 cM (i.e. < 8 cM) and will be deleted in the update. That's 28% that will be deleted.

We don't have any bubonic squirrels here yet, but I did have to change my bike route to the park because of some predatory coyotes. https://globalnews-ca.cdn.ampproject.org/v/s/globalnews.ca/news/7175663/predatory-coyotes-winnipegs-assiniboine-forest/amp/?amp_js_v=a3&amp_gsa=1&usqp=mq331AQIKAGwASDYAQE%3D#aoh=15949286295249

I don't think the testing companies will ever change to full WGS. But once WGS can phase (maybe in 5 years) and that technology comes down to consumer affordable prices (maybe 5 more years), then what they could do is take the 700,000 SNPs they have now, and replace them with the phased values from the WGS. This will make matching an order of magnitude better, probably making segments down to 7 cM valid matches. 700,000 of the most variant SNPs are perfectly fine for relative matching. No need to use the whole genome. Unfortunately, I'm a bit pessimistic that the testing companies will never do this, simply because 10 years from now, just about everyone who has DNA tested for genealogical purposes already will have, and the companies won't have a monetary incentive to update their technology. It would likely take an innovative startup, something like a GEDmatch or Borland Genetics to create that phased matching database for us.

It's still got a ways to go. It won't make a difference for relative matching until WGS can fully phase the whole genome for us. That will require accurate and very long reads that can span every two heterozygous alleles. PacBio is getting there with their new HiFi WGS, but the reads still need to be even longer while remaining accurate. And two days ago, the announcement of the first time a full human chromosome (Telomere-to-Telomere) was sequenced that included the very difficult centromere. However, they still had to manually resolve several gaps in the sequence, so they aren't quite there yet. https://phys.org/news/2020-07-scientists-human-chromosome.html

I've never been in a Zoom session with more than 25 people and it works well at that size. But how does that translate at 100 attendees, or at 500?

I really like your analysis, Paddy, and agree with most of what you say. I am intrigued by your ideas to use triangulation down to 3 cM and that Ancestry eliminate matches using a few shared match criteria rather than cM.

Blaine. I'm very surprised to see you say that. In the past, I've always passed on the article you pointed out that says WGS will increase IBD detection over microarray only by 5% to 15%, i.e., the improvement will be minimal. My own calculations got me to agree with this. Have you read something else that has now changed your mind? https://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1004144

Are they friendly enough and trusting enough to come up to you and eat out of your hand?

Bicycling is the best way to discover the beauty of where you live. There's always a surprise around every new corner.

That's a very good question. I know there are references like SNPedia that list SNP and gene properties when given the SNP name (RSID). But are there any references that give SNP and gene information that are ordered by chromosome and position?

... or maybe that.

Geneanet does not give the segment matches. But if I had to guess, it likely would be the DNA gene.

Jennifer Nani - I don't see your name in my uncle's match list at Geneanet. Maybe you're one of the X X's. Please email or DM me.

Our longest match at GEDmatch is only 5.4 cM. But I do have a real match at GEDmatch with Jonny Perl of 16.4 cM on Chr 7. I match Jonny's father and his father's half-brother on the same segment.

Kathryn - Login at the VGA website. Go to the Members Center. Then to Webinar Access. The direct link is: https://register.gotowebinar.com/register/726749004989897230

Brenda and everybody: https://www.youtube.com/watch?v=X_-q9xeOgG4

Ancestry also shows me as 100% European Jewish, which I believe is true. 23andMe show me as 99.0% Ashkenazi Jewish (not bad). But MyHeritage gives me just 83.8% Ashkenazi Jewish with the rest being strange assignments including 4.5% South Europe, 5.8% North Africa and 1.0% Eskimo/Inuit.

Thanks, Winnipeg is pretty safe from flooding. For the past 60 years, they've built storm retention ponds in all the new neighborhoods which double as beautiful lakes surrounded by parks. And the floodway around Winnipeg saves us from spring flooding.

Hopefully your area gets a good drenching soon.

So you're home from Sioux Narrows? Still, your radar is showing lots of rainfall around you. If it's that dry there, then hopefully some of it will help.

You mean none of these guys got you today? I'm just in disbelief that it can be so dry there with what we've been getting here.

You can't mean it's been dry there. It's been raining regularly here.

Beautiful! Looks so serene.

Here are FTM's instructions on a Mac. Windows is likely similar. https://support.mackiev.com/750735-How-to-partially-export-or-split-a-file-in-Family-Tree-Maker-for-Mac--

It's not easy for us. I'm lucky that on a few lines, records from the mid 1800's are being copied and indexed by researchers in a few areas of Romania and Ukraine. That allows purchase of some records with translations that can take us back 5 generations. Where that's not or not yet available, I'm stuck at 3 generations.

I believe this is the project they are writing about. Doesn't seem like there's any results yet after 3 years. https://www.jcvi.org/research/leonardo-da-vinci-dna-project

Very daring of you. Wear it proudly. You didn't have it when we were last together (RootsTech 2017)

It seems we're stuck in the same boat.

That"s a very interesting analysis. For my Jewish endogamy, 9% of my 192,000 matches are > 19 cM. But due to lack of records before the early 1800s in Eastern Europe, only 0.01% are actual relatives within my 5 generation genealogical time frame.

There's also Tablet. https://ics.hutton.ac.uk/tablet/

Whoops. Thought you said you were using them in an earlier thread, but I misread.

This article will also be interest. The writer uses NameCheap which renews at $10.29/yr. https://www.shivarweb.com/9541/network-solutions-review/

Be sure to compare renewal prices, not initial purchase prices.

Personally, I've been very happy with Netfirms (who doesn't happen to be in the linked top 10 list) for my domains and hosting for the past 15 years. Netfirms charges $15.99 annual renewal for .Com domains.
https://themegrill.com/blog/best-domain-name-registrars/

On second thought, I won't "like" your comment that opts for the more clumsy approach. ;-)

That likely can be done efficiently as well because any segment would be either paternal and/or maternal, or it would be unknown possibly with some opposite of unknown. Your matching should have no problem working with this kit and have the added advantage in some cases of being able to tell the parental side of the match.

I'm not sure what your data structures look like, but if you made a structure with 4 tracks: paternal, maternal, unknown and opposite of the unknown, you could use that for everything and not burden the user by making him keep track of all the mono and stereo kits. That will allow you to have just one master kit per person.

And another question: Are the mono sections marked paternal and maternal if they're known?

Kevin - Also, of the 7% overlapped, on average 3.5% would match and the 3.5% would be a stereo section for the other parent. How much actually ended up stereo in this case? (Hey, I think I'm actually starting to understand this!)

38% of the 18% means 7% will overlap and 11% will be new, so 38% + 11% = 49% bang on!

Just came in a couple of minutes ago, so I guess I missed most of it.

Jana Fialová Kučerová - what I've read is PacBio HiFi long reads is 99.8% accurate, which compares well to short reads which are 99.9%. Also there are no systemic errors such as repeats that are not there. All errors are random. Compare to traditional long reads that are 90% to 95% accurate and require a complex error correction step that must make all sorts of assumptions. I am very excited by the potential benefits of this new technology.

That was from this presentation: https://youtu.be/wTBTzBd3wdU

Jana Fialová Kučerová - PacBio with their new HiFi reads that are both long and accurate, now discourage polishing. They say:

I ran the MRCA Search tool a few minutes ago. It runs much faster now, taking about 150 seconds to compare 950 kits. Unfortunately I still have 0 results for my two runs as I did yesterday. I'm pretty sure that's just the way I am.

Finally, a blog post about where I am genealogically. https://www.beholdgenealogy.com/blog/?p=3470

My two runs took 8495 and 8223 seconds (over 2 hours) to complete. Neither found any common ancestors.

Ashley Bens - You're picking a needle that could have come from 1000 different haystacks and assuming it is from the one you want it to be from. The only way to "prove" an autosomal match is from a certain ancestor is to use Jim Bartlett 's "walk it back" technique, where you need a cousin at every generational level matching on the segment to show the parent it comes from at that generation. And a Y-DNA match in no way is any proof that the autosomal segment match is from that ancestor.

I have 766 kits to compare. My uncle has 933 kits to compare. If that means between 7% and 10% of my matches have uploaded their GEDCOM file, that's pretty good. So far 190 done for me and 140 for my uncle, but zero matches for either of us. All my ancestors are Jewish so I have endogamy to deal with on the DNA side, and they are all from Romania or Ukraine, so I have lack of records and ability to only go back 5 generations to deal with on the tree side.

GEDmatch is running very slow and sometimes timing out on me right now. This new feature must have got too many too excited at the same time. We're doing a Denial of Service attack on your servers.

Just had a pedigree uploaded previously. Now uploading my full tree with all living people privatized to try it out.

Thanks for the reminder to turn the page. I always forget.

It looks like a painting!

Blaine: I think your statement here is the best explanation I have ever seen as to why matches going this far back can't be done with autosomal matching.

Alona: I really don't like to think about winter and snow when I'm in the middle of summer. Corollary: I like to think about summer when I'm in the middle of winter and snow.

But those are amazing.

Dumb question. Your filename does end with .ged, doesn't it?

Amazing. Those are exactly the same pictures we took 10 years earlier.

You could make 2021 a roo calendar, and alternate each year.

Whatever they are all obviously afraid of, it must be something really terrifying.

As the mathematician, I would have concluded that the physicist did not know how to put out a fire properly, since it started again an hour later. Neither did the engineer since it started again an hour after he "put it out", and I would have then picked the best solution and pull the fire alarm and call the fire department.

I can relate. At the end of February, leaving for vacation, I left my computer running an assembly. When I came back two weeks later, the SSD failed and I had a blue screen. Here's the story: https://www.beholdgenealogy.com/blog/?p=3269 One recommendation: For any Windows computer, get Diskeeper by Condusiv, for background defragging and SSD management.

Whoa! Didn't realize this. Terrible. She's one of the most fantastic singer/songwriter/fiddlers anywhere. New page seems to be catching on.

That's just the sound of a satisfied squirrel who is calmly eating. It's nothing like the screech of an antsy squirrel arguing or fighting or escaping or attacking. I hear the latter more often.

Tim Shaw - Unfortunately, WGS tests today are far from perfect. Short reads cannot span repeats so they don't assemble or align well and can't be relied on to correctly produce the novel variants needed for the polymorphism comparisons you're talking about. Whereas long reads today are too error filled, like up to 10%. And although they are long enough to span repeats, the algorithms to correct the errors and align or assemble them will never be able to be made to produce results good enough for Genetic genealogy. 6% of my long read aligned SNPs were wrong when comparing them to my microarray and short read results. So doing a WGS test and saving it is not a good recommendation with what's out there today. We need to wait for a better WGS test such as PacBio's HiFi test, which has accurate long reads. That technology is only a year old and analysts are still developing the assembly algorithms that will take advantage of it. Until the HiFi test or some other innovation like it becomes commercially available, I wouldn't recommend WGS today, because today's tests will never be better than the microarray tests we already have for genealogy.

Can be even better if there's also koalas in the trees.

I used to watch John de Lancie when he played Eugene Bradford on the daytime soap opera: Days of Our Lives. When he first appeared as Q on Deep Space Nine, I shouted "Eugene" and still call him that. Love his character.

Maybe if you start feeding them like Alona.

Actually, that's a great idea. I'm not very handy with wood, but maybe I can find something similar that would fit on my fence.

p.s. My bait of a few walnuts at the same place on my deck each day is being taken, but I haven't caught the little guy in action snatching it yet.

Good suggestion! A segment age probability calculator would be a very simple tool for Jonny Perl to add to his DNA Painter site. The calculations are almost trivial compared to Shared cM and WATO.

Blaine - I was thinking about it all night. And then when I saw your post here this morning, I just had to do it. I'm still in my pajamas.

Graham Hart - No. I see it as Kevin is properly working bottom up and seeing how many generations back the segment will go without a recombination creating it. If you went higher, then you'd find the, say, 7 cM was now a 5 cM segment on one line, and a 2 cM segment on another other line. So anyone matching that full 7 cM segment would have to be from a generation up to that recombination. Kevin describes that quite well in his article.

I just posted an article comparing Kevin's estimates to Speed and Baldings: https://www.beholdgenealogy.com/blog/?p=3420

Iain - Oh I see now. You were talking about the free long read offer I mentioned in the other thread.

I am far from an expert on WSL, just dabbling so far with Ubuntu in WSL1 to use some genetics tools. But what I heard at Microsoft Build last month was that WSL2 is fantastic. It has full access to the Windows OS and filesystem and multiple, even different versions of Linux, can be run at once. For genomics applications, the performance improvements are the best reason to switch to using WSL2.

Iain Kennedy - Not sure what you mean. The Sequel II is required to do PacBio HiFi.

I'd like to find we're related so I can go visit.

Someone in Vanuatu!?

They did a Long Reads test for me in June 2019 and I have my results.

Patrick - I went through PacBio's Certified Service Providers list that offer the Sequel II System. I recognized the name DNA Link from somewhere, and was impressed by their site. And their landing page from the PacBio list caught my attention right away when it said: "20th Anniversary Limited Offer. Free long-read NGS analysis" DBA Link's headquarters is in Korea, but they also have offices in San Diego and the Netherlands. https://www.pacb.com/products-and-services/service-providers/?fwp_csp_instruments=sequel-ii-system&fwp_sort=preserve

Well, just two months ago, they offered everything. So it's not been that long.

DNA Link Sequencing Lab seems to offer it if you have a specific project. They write: "Whole genome sequencing for de novo assembly using SMRT Sequencing is now the gold-standard for generating contiguous, highly accurate reference genomes across all species. With megabase-size contig N50s, consensus accuracies >99%, and tools for phasing haplotypes, PacBio assemblies capture undetected SNPs, fully intact genes, and regulatory regions embedded in complex structures that fragmented draft genomes often miss." https://www.dnalinkseqlab.com/pacbio-sequel-rsii/

I also found this very informative: https://programs.pacificbiosciences.com/l/1652/2020-03-20/41r14g

Robin P Meyer - One page request from a tool is exactly the same as one page request a person would make. Ancestry's servers have a certain capacity, guessing 1000 page requests a second. Once they reach the capacity, the requests will slow down and start to fail. So 1000 requests per second can handle 10,000 people doing 1 request every 10 seconds. But it can only handle 10 bots doing 100 requests a second. Slowing a bot down to 1 request every 10 seconds means it would take over 24 hours to do 10,000 requests. So it's feasible for some types of analysis.

Evert-Jan - I think Ancestry might be having server load problems just from their tens of millions of subscribers. They likely started contemplating additional capacity which is expensive. Load shedding is a cheaper way to accomplish the same thing, but hurts users who use the extra services. It"s a trade off that Ancestry thought about and finally decided to do. They may be waiting until all bots are halted to then see what the server load is down to, and then will reevaluate if they still need additional capacity.

RainMan Jim - GA was by no means the only bot running, and likely wasn't one of the heaviest users among the bots either.

My thinking is that the automated tools of the various companies were starting to put a significant load on Ancestry's servers. The past few months, Ancestry's subscribing users have been faced with poor response times and many request failures. I don't think Ancestry would have done anything if this hadn't been the case. They had been silently allowing all the bots for years. Don't think they didn't know about them. Unfortunately and ironically, it's the success of the bots that proved their own undoing. Each one can make requests thousands of times faster than a user at his keyboard. Okay, hundreds then, if you guys are throttling. And your success with thousands of users leads to hundreds of thousands of requests, likely swamping their servers. They can tell from their logs that they know it's you. And they've taken the logical action unfortunately. They wanted a quick solution so they used their terms of service along with cease and desist with their real goal to just get all those bots to stop. It wasn't about privacy nor copying data. It was all server load. At least that's my analysis.

It's more of an art now than a science, and it's currently expensive. But maybe a few years from now. https://www.theatlantic.com/science/archive/2019/03/dna-tests-for-envelopes-have-a-price/583636/

Save it for the day you can reliably get it DNA tested!

Do you know if that's a lossless or lossy CRAM file? https://www.uppmax.uu.se/support/user-guides/using-cram-to-compress-bam-files/

Was the blue one your sibling? If so, you may be matching with your sibling on the short segment through the parent that triangulates, and you are matching your sibling on the rest of the segment through your other parent.

Randy - yes they do "polishing" (using short reads to "correct" long reads) sometimes several times in both alignment and assembly. But I'm worried that the polishing could "correct" some long values to be incorrect, simply when the short reads get aligned incorrectly, which to me is very possible in de novo assembly. I haven't seen any papers evaluating the correctness of polishing. Polishing sounds like something you'd like to do until you really think about it.

Question: I have a large tree with 5000 people, including about 1500 who I am genetically related to. I have taken a MyHeritage DNA test and have 17000 matches. My uncle who is in my tree has 15000 DNA matches. But neither of us have any theories. For security reasons, I have manually in Family Tree Builder set all of my living people private, so that they only show up as "Unknown" in MyHeritage, i.e. no surnames. Could it be that this privacy setting is preventing your algorithms from finding any theories for me or my uncle?

Hi from Winnipeg, Canada

Our unions at Manitoba Hydro negotiated a 4-5 work week about 25 years ago. Every 2nd week where there wasn't a civic holiday, we'd get a Monday off. In order to work the same amount of time, the agreement was that we'd work each day from 8:00 to 4:40 instead of just to 4:00. I loved that arrangement.

Helen - NCBI's dbSNP database now in Build 153 (August 2019) has 695 million human RS records, including 552 million with population frequency data. That's more than 20% of our genome where SNPs have been found. 12 million RS records were added since Build 152 (December 2018), so more are being found all the time.

The 1000 Genomes project found about 85 million SNPs for all the people. But one person's genome typically differs from the human reference genome at 4 to 5 million SNP positions. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4750478/

Oh. I'm surprised you hadn't already tested everywhere before this year. All the better for Humpty Dumpty.

"An assembly is a hypothesis." - Karen Miga

Judy - I thought you would.

Surely you must have tested with FTDNA before. If so, why are you testing again?

Judy - I've never seen one wild. Apparently possums eat anything. And North American possums will eat baby squirrels. https://thesquirrelboard.com/forums/showthread.php?37122-Are-Possums-dangerous-to-squirrels

Thomas - Your article says about 2 years to go. I don't like it when they need to polish and then do it twice. I think polishing may clean up some errors but introduce others. The short reads may be more accurate, but they can easily be aligned in the wrong place. So I don't fully trust polishing yet. I'm hoping one day for a single run assembly of all 46 chromosomes from a single long reads test. Maybe 5 years?

Not sure if I can like your possums yet. They look like outlaw cousins of our squirrels, and that's bad when you consider Judy's feelings about squirrels.

Amit - Hi François. I'm interested in what they have to say about the generation of the first human telomere-to-telomere assemblies of whole chromosomes. I hadn't heard of that being done before. Does the article state what tools and/or pipelines are being used to accomplishing this?

François - Thank you for that interesting article. It does seem that the current method of aligning reads to a human reference is inadequate. The article does state that long reads with de novo assembly is needed for full reconstruction. It does not, however, address the fact that long reads with current technology have a high error rate, so specialized algorithms are needed to compensate for that.

If you could put a thin black box around each chromosome, then the background can be white. I'm thinking of something like this. Then it's obvious where you most match.

Hmm. You have black for no match, and a dark color for full match but a light color for half match. That's a bit confusing. Maybe white for no match would be better? ==> the darker the color, the more you match.

Jason- the best list of microarray file formats that I've seen is this one by Randy Harr. https://h600.org/wiki/Microarray+File+Formats

I don't know how you do it Kevin. I can be that ambitious with my programming work sometimes. But progress is never so fast for me.

Yes. Very interesting. Just for comparison, I'm K1a1b1a (Ukraine) and my uncle (i.e. my father's mother's line) is H3w (Romania). Many of both our mtDNA matches at FTDNA have names that are likely Ashkenazi.

I like and use partner for unmarried parents. Marriages and partners are "together" only until they separate. Genealogy Software allows you to specify both the start and end of marriages and partnerships.

I live in Canada. All are from different countries. 5 from USA, 1 from Argentina and 4 from various countries in Europe.

Very interesting. Do you know the background of the surnames that at least on the surface, seem to have been taken from town names, such as Focsaner and Zvorestineau?

This creaping up phenomena is very interesting to hear about. Ancestry raised me from 98% to 100% a year ago which is where I think I should be. When I first took their test, they had me at 86%. The other companies have me from 84% (My Heritage), 92% FTDNA, to 99% (23andMe)

Pierre - I use MailWasher for my email.

Shhhh. They're listening and you don't want them to notice you. No matter what you're trying to do to be a good citizen, you're not compliant to their T&Cs as far as they're concerned.

I have a few bunnies around me now, and this is the best I can do.

Dean - I'd definitely pay that today for simple plug-and-play spam protection. But 15 years ago when I was paying $3 a month for web hosting, I couldn't justify paying the $5 a month for a service that then had some good free alternatives.

Banai - The WordPress settings page had two site links and changing those got the sites working, but as you say there were a few physical links I had to find and fix. Your move the registration page idea is a good one. Unfortunately my registration page is part of my wp-login.php page which I'm sure the bots look for. Fortunately, my spam registrations are all being prevented, now 21 since I got it working 4 hours ago.

Ah yes, Akismet is a good one. That's likely what most WordPress people use. It was around with my version of WordPress but I remember having some problems getting it working. And I believe they wanted to charge me a monthly fee back then which I didn't want to pay. So I found other solutions.

It's hard to get close enough to rabbits and squirrels to take good pictures. How long did you have to stay perfectly still for them to get close?

Excellent point, Pierre. Noone really needs permission to write something on their own about any product. If it's a complementary article that respects copyright, the product owner will be quite pleased to get the free publicity, and wouldn't mind if a few images from their site are used. So either this person has something fishy in mind, or they just don't have any idea how to set up a web service site.

Thanks everyone. Like you say: 1. The lack of anything of quality on their site, 2. Their unwillingness to state who they are, and 3. Your 0 to 4 unanimous vote seals the deal. Thomas, yes I get a lot of spam as well, especially for website makeovers and SEO. I got one the other day from someone seemingly wanting to buy one of the extra domains I own. But this one seems a bit different being custom written for me with a genealogy connection.

I am not very generous in handing out "likes". But your pics of what you've turned into everybody's koala and roo friends always bring a smile to my face and gets a like. I'm sure you've got 300 of my last 600 likes.

That"s everything in their terms of service that mentions law enforcement.

Their terms of service state: "By accepting these Terms, you also acknowledge that you understand that nothing prevents a court from issuing a warrant or other legal request for information stored on this site, and in such cases, Borland Genetics will comply with any such warrant or request in good faith to the extent required by law. If you do not consent to such uses described herein, please do not upload your DNA data to this site. By providing your private DNA information to Borland Genetics, you are opting to permit any and all such uses of your data as described herein, including use for law enforcement purposes."

Bob: You can subscribe to a non-recurring subsciption for one-month or for one-year which is like a one-time unlock cost. Or you can subscribe to a recurring subscription and pay monthly or annually.

Well, it was either that or when we both were dressed as Klingons at a Star Trek conference.

When I showed up as a DNA match of yours on our common ancestral line in Kazakhstan. (Sorry, but I never change my Facebook status)

The "Father's side" is a great feature, but only available for those fortunate enough to have been able to DNA test their parent.

I only have the Ancestry Canadian subscription, so almost all of the hints are to documents I can't look at.

Hi from Winnipeg, Manitoba, Canada

Also, I learned something very important. I'll likely be buying a bag of walnuts next time I go shopping. Without building a full ninja course, any ideas on what I should try?

I think my squirrels Hopper and Speedy must be training for this.

Yes. I understand. But there are nuances. e.g. You said: "then the same length Phoenix segment has to be from his mother that he shares with you ie your grandmother." Yes, that segment will be my uncle's from my grandmother. But it need not be a single segment in my grandmother. Part of it may be from my grandmother's father and part from my grandmother's mother if there was a recombination there when passed to my uncle. If so, then part of that segment in my uncle will match some people on his father's side and the other part will match some people on his mother's side. Painting my uncle's chromosomes is needed to determine the assignments of these Dark Side segments. Most of my segment matches with known relatives which gave me the information to paint my own chromosomes was at 23andMe. My uncle was tested at Family Tree DNA. 23andMe doesn't accept uploads so I don't have that information to paint his chromosomes. So for the time being, I can't make use of the reverse Dark Size information.

I signed up back then for my free realignment. Never got it. Never got any notification about it. I don't think they did it for anybody.

Zero. I only go back to 3rd great grandparents. There just aren't records any further back than early 1800s in Romania and Ukraine.

Most of the tools are free. There are a few advanced tools that require a subscription.

Jason: First, thank you very much for very being patient with me with your explanation. I am trying to work through what you mean and see your point. But to me the problem is that the result of Darkside with my uncle will give segments that may be his father and may be his mother. So it can't be used to create a kit for either his father or his mother. e.g. All Chr 1 my uncle and I match. Say that Chr came from his father. Then his Darkside is his mother. All Chr 2 my uncle and I match, but that Chr came from his mother. His Darkside for Chr 2 is his father. The Darkside output would be a combination of his father and his mother. The reason why it doesn't work in reverse for Darkside is because I share DNA with both my uncle's parents. But Darkside works for me because my uncle doesn't share DNA with both my parents and I can use it to determine my mother. But then I'm sure you know this and were trying to explain it to me because in your original comment, you said: "you would still need to determine if the segments were his mother or his father". So it's not usable in the Borland Genetics system because it doesn't represent a specific person yet.

Great couple! Great family! 45 more!

Jason. Yes, if we share, then we got the same one. But we were talking about what chromosome my uncle might have if he and I do not share on the segment. He could have got it from any of his other three grandparents that did not pass down the segment to me through my father.

Jason - No. It might be his opposite parent, or it can be his same parent but the other chromosome. None-the-less, this analysis still can be done to get my uncle's segments that I don't match. Then if I do a DNA Painter analysis of my uncle, I can add his segments to the other ancestors that I've done for myself through DNA Painter.

Ah yes, Season 2 episode 1. I was trying to verify that but RR took down their season 1 episodes. See below starting about at 32:24 https://www.byutv.org/player/6eedba2c-1898-476e-a389-f08696c6bb89/relative-race-season-2-episode-1

Yes Al, it would have been nice to make this an annual event. I guess we could have gone to a patio restaurant, but my family's still not wanting to do that yet. Wonderful that you were able to have a great birthday and enjoy your granddaughter.

I'm going to be posting an article on my blog about Borland Genetics in the next few days.

Yes Kevin. I do understand that. But for someone like me who has only the raw data for myself and my uncle and only distant matches from your database, I can't get very far. What I'm saying is if you add segment match information (without the raw data) for me to known close relatives and for my uncle to the same close relatives, that will allow in common and not in common parts of my and my uncle's raw data to be assigned to ancestors farther back than can be done with the raw data alone, sort of in a DNA painter fashion.

Do you see what I'm thinking, Kevin? Borland Genetics should also segment matches with on each grandparent's side for every person with raw data. This could really enhance BG's analysis possibilities. For example, I only have raw data for myself and my uncle. So off the bat I can create a bit of my father and my mother. But if I add my uncle's paternal and maternal segment matches, and compare with my grandparental segment matches, well then now I can assign and extract grandparents from my and my uncle's raw DNA.

You see, I don't think you need to upload the phased kit to GEDmatch to get the matches. Accepting user-supplied "phased" matches would be just as good.

So wouldn't those segment matches combined with the raw data for the 3 siblings be able to give just about the best Visual Phasing possible in one automated step?

What if a person was able to supply a set of segment matches (chr, start, end) with each child to known relatives of each grandparent?

Kevin: Can't the equivalent of Visual Phasing be done with the currently available Borland Genetics tools?

I believe the majority of the contents of the book is under the "Read the book" tab. Algorithms I've never seen described before, and described so well! But they still would like it if you buy the physical book, and use it for teaching a course.

I mailed a kit from Canada to Dante Italy on April 18. They received it on May 7.

Alona: We've got some rabbits here, but I haven't figured out how best to tell them apart. What's your technique?

I'm thinking that they don't know the person buried there. In 1986, my father and I went to the cemetery his step-father was buried 40 years earlier. But the family couldn't afford a headstone back then. We had one made, but their records didn't have where he was buried. They had about 16 unknown burials. They said placing his memorial on one of the unknown plots would still honor him. So we picked one of them and had the headstone placed there.

During DNA week a few weeks ago, they offered their lifetime subscription for $200 (reg price $700). So then their WGS and lifetime subscription was $500.

Well done! Enjoyed the guided tour and walk through Leiden, along with the interesting pilgrim information.

Hi from Winnipeg, Manitoba, Canada. I visited Leiden 6 years ago. Beautiful city!

Hmm. That's good to know. No one's complained about the sound from mine, but maybe I should do a test too.

Several years ago, I needed a Webcam. Pat Richey-Erickson recommended the Logitech C920 Pro. It's a Webcam with an excellent microphone in it. I love it. It's small. Great quality. And doesn't give feedback because it's aimed at and picks up you, not your speakers. https://www.amazon.com/Logitech-C920-Pro-Webcam-Black/dp/B00829D0GM/ref=mp_s_a_1_3

We likely triangulate with many of our shared friends on Facebook. They do allow us to check that. But they don't give us segment data or ethnicities.

Kyle - I'm not familiar with Clinvar, but my direct analysis of my filtered VCF and my raw (unfiltered) gVCF files indicated that 0.08% of my filtered VCF variants were wrong and 0.13% of my raw gVCF values were wrong. That's not too bad, but the problem is that my filtered VCF excluded 0.76% of true variants. That is a rather high value, meaning the filtered VCF is missing a lot of variants. I think it's very dangerous to miss an important variant you might have. The raw VCF only excluded 0.09% of true variants, but that's the tradeoff: Lower Type I errors means higher Type II errors. The more aggressively you filter, the fewer incorrect variants you'll have, but you'll also filter out many true variants. https://www.beholdgenealogy.com/blog/?p=3021

Kyle - I disagree that there are a lot of miscalls in chip (genotyped) data. The criteria for the company treating them as a no-call eliminates almost all false positives. So I'm unsure what makes you think they produce "an enormous amount of erroneous data". No-calls are not errors. My Ancestry DNA data does not include InDels. For the SNPs that chips give you values for, the data is very good. WGS can't and doesn't do any better as far as those SNPs are concerned. What WGS does is give you good values for the 6 million SNPs you have that chip tests don't give you. Have you ever compared two WGS tests done on one person? The amount of differences are astounding and will make you question how good WGS is.

Great pic. I pulled my phone out for a video but I was too late.

Kyle: Interesting article. I like your point of view and some of the comparisons you've made that I haven't seen elsewhere. I have a few comments which are meant as constructive criticisms so I hope you take them that way: 1. In all cases, you call all the chip tests "Not as good as Whole Genome Sequencing". You don't explain what your "good" is. I would call "good" to be accurate SNPs for the variants you're looking for and not having inaccurate SNPs at those positions. I would not necessarily say WGS is better at that because there are many mistakes in reads and alignments and the rest of the analysis that gives you your WGS results. 2. Chip tests give you way more than 0.02% of your variants. It is estimated that there are about 30 million SNPs in all people and one person may have 6 million, with any two people averaging 3 million SNPs different from each other. A chip test will likely give you 300,000 of your 6 million variants, or about 5%. 3. You mention insertion/deletion variants like BRCA are often not correctly called by chip tests and you link to an article about that. The other side of the story you don't mention is that 23andMe's BRCA analysis has helped people, such as genealogist Lara Diamond. https://thelayersprojectmagazine.com/stumbled-upon-brca-results-saved-life/ 4. 23andMe is the only company that includes InDels in their raw data file. Yes, their accuracy may be low, but as long as that is realized, it's still better to have that data to analyze than not to have it. Getting InDel data from WGS results is not easy nor fun. I question how good the WGS alignment procedures are at aligning around InDels accurately and I ask the same question re 23andMe's chip test as to how they can determine an InDel is at that position. I expect once de novo assembly becomes more available, we'll get a better handle on our InDels. 5. I do not like your mentioning of imputation as an easy way to get some extra SNPs. Imputation uses population averages to assume what your SNPs are. Making that assumption for important health SNPs can result in horrible mistakes. I'd like to see an article that assesses the ratio of false positives and true negatives in the imputation of health-related SNPs. I would bet the accuracy could be well under 50% for rare variants, something that no-one would want to rely on. The 99% accuracy of actual readings is much more assuring. 6. Through the article, you talk about miscalls and even "bad call". Most chip companies replace SNP results they are cannot determine with a no-call (i.e. something like a double dash). There could be as many as 3% no-calls. More than that and the company will likely want to resample you. But mistakes on what the chip companies assign values to are quite rare. Comparing my 5 tests and ignoring no-calls, I found only 665 of the 1.4 million different positions had differing values, or 0.05%. https://www.beholdgenealogy.com/blog/?p=2700. I don't think WGS tests are as good as that. Louis

Squirrel Olympics

Dana - Thanks for mentioning that you as well have instances of one instead of two ancestors. That tells me that mine isn't an isolated incident.

Dan: He has no dates on Joseph or on Esther.

Interesting in the first video that Speedy didn't go over the corner post, but went around it.

Craig: Your answer sounded promising. But then I looked at my cousin's tree, and he has just a minimal tree and only the name of each of his grandparents and no dates or places. So I don't see why his "joseph girman" would not match up to my "Joseph Girman" when his "esther goretzky" does match to my "Esther Baila Goretsky" (with the last names spelled differently).

Thanks Paula: Unfortunately, I don't have edit capability on my cousin's account so when I look at his tree the Edit Relationships box isn't available and I can't see whether Frank's father is marked as biological or not. When I look at my cousin's tree, this information shows under his father Frank. If his tree had Frank as not biological, would it show here?

Do you mean you actually got them all into a single box?! Hats off to you if you did.

Not sure why, but I can't seem to find any roo food where I am.

Unbelievable job! But for the "that might be an important family history item" freak, what did you do with all those loose items and papers? Surely you put them somewhere so that you later can go through and make the keep/throw decision carefully.

See also Lara Diamond's article about doing that at GEDmatch. https://larasgenealogy.blogspot.com/2017/03/a-technique-for-endogamous-dna-using.html

Of course.

Poor Marmite. We remember her well.

Squirrels are so hard to photo because they don't stay still. Good catch!

Brenda - Multiplication is always done before addition. If they said 20, they're wrong.

You were too fast for me to correct. . Forgot the times at the end. 15.

12

Thanks for the very informative explanation. When you say "strand", are you actually meaning one of a chromosome pair which is a "chromatid"? One chromatid comes from the father and one from the mother. Isn't it the chromatids that we want to phase? Whereas each chromatid is made up of two strands: forward and reverse where each position is either AT or CG. Also, I don't think full phasing is possible yet for the human genome. The long read segments are still not long enough to connect over all the regions where the two chromatids have the same value.

Wow! "Shasta is an in-memory computing-driven algorithm that can now help complete a de novo (new, never before processed) human genome assembly in under six hours, the authors say, for an average cost of $70 per sample." Shasta is available here: https://github.com/chanzuckerberg/shasta

They have so much time, they could do genealogy.

Too many links. Not enough time in the world.

Alona - You nose your koalas very well. I'm still trying to figure out a way to identify our squirrels. They just won't sit still long enough to get a good look at them.

In Canada, we drink maple syrup straight. https://www.bardown.com/vasek-pospisil-sipping-maple-syrup-from-the-bottle-is-the-most-canadian-thing-we-ve-seen-1.1440171

https://www.forbes.com/sites/jrose/2019/03/21/how-much-do-youtubers-really-make/#629942517d2b

That seems to be a trend also with genealogy companies reducing their affiliate offerings. In the last few years, a number of the affiliates I've been using have dropped their programs. So much so, that I decided it wasn't worth my effort to do so any more. Of course I was nowhere as reliant on the revenue stream for my income as you are. I wasn't making more than $1000 a year from my affiliate commissions.

Nebula's FAQ says they do 150 bp paired-end read sequencing using high-throughput MGI DNBSEQ-T7 DNA sequencing machines. They align to hg38 using a GATK-based pipeline. Dante switched to Illumina NovaSeq 6000 systems when they opened their new sequencing centre in Italy last year. The machines are probably quite comparable. I would expect short read tests from the two companies would give similar results. There likely would be more variation from the quality of the sample than from the difference in the machines. The analysis pipelines could make a difference as well. I don't know what Dante currently uses for its short read pipeline.

Yes, that's correct. And I believe it included return airfare from Toronto to London.

I suspect that the background noise for Ashkenazi is about 40 cM, which is why the clustering works for me for 50 cM but breaks down at 40 cM. That seems to be as good as Timber can do.

But 40 cM is too small

This is my 100% Ashkenazi endogamy at 50 cM and it looks fine. I was able to identify my 4 grandparent clusters, e.g. FM = my father's mother. On my mother's side, I have a known 1st cousin and a 1st cousin once removed between the MM and MF blocks that span both blocks.

Only so much room up there in that grey mass of ours. In with the new, out with the old.

Thanks Linda. I'm a Windows guy, but I've heard good things about MacFamily Tree.

Linda - Are you able to record all the Chinese names of your ancestors in Chinese on Ancestry, or do you have to type the names in English?

Question: Have you been able to access any resources from China and trace even further back?

Hi from Winnipeg. Looking forward to hearing about how to do Chinese genealogy.

Randy - I'll send you them, but I don't think this will help. Don't non-closely related people have mostly different indels. And other people may have single SNPs at my indel positions.

I haven't attempted the Microsoft Store mountain myself yet. Definitely best to KISS.

Not sure what method you're planning to use, but I've been including a Comodo Code Signing Certificate for years in my software and install routines. They will work on any version of Windows. You can get them now for less than $80 a year. They used to be several hundred dollars per year.

Weren't the big fires to the South and East of you? You walked North and that all looks pretty.

Umm... Sparky and McCoy

You spoil them sooooo much, Alona.

What would a roo prefer: pellets or apples?

But my Short Reads file seems to be CC. I'm nowhere near being an expert at reading these things, so I can't say I'm necessarily interpreting this correctly.

Randy - I do have my list of 3408 insertions (II), 1345 deletions (DD) and 47 ins-dels (DI) that 23andMe provided me. The images are what Tablet shows for my Long Reads BAM file and then for my Short Reads BAM file at the position 25773865 on Chr 3 that you gave in your example. WGS_Extract gives me CC, but 23andMe tells me it's a deletion (DD). Here's my Long Reads file. It looks like a deletion on one chromosome because half the reads at that position are *.

Nathan - I ordered an upgrade a few weeks ago. They are sending me a new test kit because my long reads sample was all used up and my short reads sample is too old. My understanding was that they would be assembling your genome from the long and short reads without respect to any reference genome, which is what I want. Dante has updated their pages and many of the details have been taken out from their Whole Genome H Upgrade page (and I see they now accept PayPal!!), but they used to say the following for the Premium to WGH upgrade: • First Hybrid Whole Genome combining Short + Long Reads • 130X for the Exome, 30X for the rest of the Genome, 15X Genome Long Reads • Hybrid Genome assembly using Short and Long Reads • Raw Data available free of charge(FASTQ, BAM and VCF) • Sequenced in our own European Sequencing Center 8 week turnaround time

Randy - Including InDels is still a problem for genetic genealogy. InDels are not InDels for everybody. Some people have normal SNPs in those positions. So reporting an InDel as a SNP will give an incorrect result for genealogy purposes. Somehow Thomas should change Extract23 to at least detect if an InDel is at that position and either mark it one of, II, DD, DI as 23andMe does, or give no result at that position, or mark it as a no call.

https://www.youtube.com/watch?v=a9xAKttWgP4
p.s. John Conway passed away a couple of days ago at the age of 82.

Great catch, Judy. They don't stop moving for long. Those claws can climb anything, even up the side of my stucco-coated house.

I started in Fortran. Did two big projects at Hydro, first in PL/I, second in Pascal. Never did much in COBOL which was for our business software. At university, APL was by far my favorite. Now I program in Delphi which is Pascal-based.

Winnipeg, Manitoba, Canada.

I've only had 3 residences:
The home I grew up in.
The apartment we rented after I married.
The home we bought and still live in.

Kevin - You're correct. The standard microarray testing companies use the same technology. The differences between 0.2% and 0.5% are likely random variation. It doesn't mean that one company necessarily has fewer errors than another.

It should be 30 total for both alleles, and average 15 for each allele if 30x is at that position.

So sorry to hear this. Fritz was our driver M22-1979 and I heard he did many of the first and last tours every year for many years. It is nice to see he lived to the ripe age of 89. Must have had a good life. Is there any info online about Fritz and his life?

Randy - I never took the Nat Geo test. Maybe Ann Turner might have it.

Marko - Yes. I'll continue to respect your right to privacy.

Marko: More about c): Doing a study of GEDmatch 1:1 would introduce yet another variable being the matching algorithm. I've got enough variables already with test types and alignment algorithms. Matching algorithm analysis is a whole other beast.

Marko: b): No one test will give totally accurate results. And yes, some tests and some reference genomes will give more accuracy than others, as well alignment algorithms, de novo assembly algorithms. My goal is not to try to find the best of these, because there likely isn't anything that is best at everything. I've taken a few WGS tests (with one more and a de novo assembly coming) and by comparing them and various analyses of them to each other and to my 5 microarray tests, I next will be using them to correct each other and estimate error rates of the tests and the algorithms. I have enough variety already to do this now. If I was someone with just one WGS test, then yes, I'd likely want from hg19 to hs37d5. But that's not my goal here. I do have one build 38 BAM which I would think is again better than hs37d5. With all that said, you make the hg19 versus hs37d5 comparison sound interesting. Have you posted your finding anywhere, or has anyone else done something similar? c) Yes, I do realize INDELs are not considered. I'll make note of any I have in my microarray results when I do my comparisons. d) Thanks for the Y-DNA info. I did Y-500 at FTDNA a couple of years ago to help the Ashkenazi study. They've got experts doing the analysis and I'm happy to leave them toit.

Randy - The WGS Extract Manual is superbly done! Same for your Bioinformatic Tools on Windows 10 Quick and Dirty document that you refer to on page 1 of the WGS manual. May I provide links to both of them in my article? Also, may I quote the first two lines of the 4th paragraph: "The tool is the concept of Marko Bauer and … heavily modified."

The problem is definitely something in the version of the mpileup code that Marko used preventing Long Reads from working. Using the -B option fixes it. Using the latest version of htslib might fix it as well and if it does, that's the better solution. See my post about working with the new version of WGS Extract: http://www.beholdgenealogy.com/blog/?p=3294

I think you got it, Randy. Using this post: https://github.com/samtools/samtools/issues/778 - I took their suggestion of including the -B option, and I was able to run WGS Extract successfully on my Dante Long Reads file. But that post also says that using the latest version of htslib as you mention also fixes the problem. That would be a better solution since we don't want to not do that quality check. I've now run my Short Reads test both with and without the -B option and I'll be doing a comparison.

Based on my observation that the Long Reads files weren't being built, I searched the web for some reason why mpileup might not be working with Nanopore WGS. I found this article and based on it, I tried adding the -B option to mpileup. It seems to be working now. I'll report back once I've tested it with my 2 Long Reads and 1 Short Reads file. Hopefully this is the fix because it's a simple one if it is. https://github.com/samtools/samtools/issues/778

I have discovered the "temp" directory in WGS Extract is where the intermediate results are stored while the program executes. As mpileup runs, a file called temp_austosomes_raw_vcf.gz is built up. For my Short Reads file, after about 3 minutes, that file got to 3092 KB in size. I stopped the run and decompressed the file, and it had 69,927 lines in it that had extracts up to chr 1, position 92,101,761. For either of my two Long Reads file, after about 15 minutes, that gz file was still at 0 KB. There was nothing in it. My suspicion is that the mpileup program does not work on Long Reads files. :-(

Okay. Then I'll just have to call you the mastermind behind the idea of it :-)

Thanks Thomas. I was running the first minimap2 assembly step on my own computer and hoped it to finish before I got back from a 14 day holiday a month ago, but my SSD failed while it was running. I'll send you an email about doing the assembly. http://www.beholdgenealogy.com/blog/?p=3269

Hi Thomas. Yes, that was the BWA run that I did myself. (Took 5 days). You did a minimap2 run for me at YSEQ three months ago, and I was next going to run WGS Extract against it. I wanted to compare to see how different the results from the two algorithms are so that I could get an idea of how much the error rate improves with a better algorithm. p.s. I am still interested in the de novo assembly we had emailed about on January 4. I have ordered one from Dante (see above response to Nathan) and I've asked them for their pipeline. This is all for personal learning purposes and then education of others through my blog posts.

Nathan: I'm hoping to look at and compare the Long Reads vs my Short Read vs my standard DNA tests to try to determine error rates and improve all the readings through comparison. I've also ordered a Dante de novo assembly (their Whole Genome Hybrid) that combines both long and short reads and hope to see if that produces more accurate results.

It's going up to 6 C today. Some of the 20 cM of snow we got in the past two days might start to melt.

Hurt? Grow? What it definitely will do is cause change. Conferences won't be needed for learning any more. So that leaves vendor exhibitions and in-person networking. And I'm sorry to say that I think it's going to be very difficult to be profitable through providing online talks now. There's just too many free or inexpensive recorded talks available already. The only hope is completely new material.

Have a great birthday, Helen 🎈🎂

This is after shoveling this afternoon. Still 24 more hours of snow to fall.

Since we'll likely get more than 15 cM, that means our 2 biggest snowfalls this winter were October and April. Here's my daughters this morning scraping the ice and snow off the car windows:

Al - The ultimate outcome will be a big increase in 2021 of both babies and divorces.

Ha ha! Here is the Canadian version.

No, but this is news from today: https://www.cnn.com/2020/03/31/world/antarctica-heat-wave-intl-scli-scn/index.html

R S Vivs Laliberte - I only had 6 data points to work with. I did average out the 4 female results together and the 2 male results together, but I don't think you can say conclusively that one is higher than the other in the 10 to 20 cM range.

I did a study 3 years ago seeing what sized segments are also a match in one of a person's parents. For the X chromosome, it was really bad, with as many as 20% of those 15 cM not also being a match in a parent. http://www.beholdgenealogy.com/blog/?p=2003

That was an excellent presentation with some material I had never seen before. Alex's presentation style reminded me so much of Blaine Bettinger. I can't think of a better compliment than that.

Happy birthday, Alona!

David: When you run Person A versus Person B, all results are in a single spreadsheet file that contains all chromosomes on one page, and the people information on a second page. When you run Person A versus all the people in Folder B and select "Combine all results", then DMT produces 23 files, one for each chromosome and also a separate people file. It does this because you may be combining dozens of A versus B matches together and that can result in too large a file if it were put together. DMT has always done that. You just are obviously now running Person A vs Person B whereas previously you did a Combine All Results run. The coloring has changed starting in Version 3 since DMT is now set up to determine your ancestral line on each segment. You can get DMT to do this by entering the MRCAs (Most Recent Common Ancestors) for the DNA relatives that you know into the People file. The only color you'll have is purple if you haven't entered any MRCAs. I highly recommend you read the complete DMT help file. It explains everything that I've been telling you here and more. Just press the Help button in the middle of the DMT window.

David. You enter Person A's one-to-many file in the File A box. You enter Person B's one-to-many file in the File B box. You save any one-to-one files with the Save GEDmatch 1-1 button into the folder where File A is. DMT will add the matches of any one-to-one files that are in Folder A to Person A's matches. You don't have to manually do that. The person you enter in File A is the person whose genome and matches you want analyzed. Usually that will be yourself. If you use your cousin as Person A, then you'll be analyzing your cousin's genome, not yours.

Picard of course. But also Zoey's Extraordinary Playlist. It's not a musical but if you love musicals, this one's for you. Most unique concept ever. You'll laugh. You'll cry. See it from episode 1 or you won't understand it.

David: You still must enter Person A's segment match file as File A. Then put the One-to-one file in the same directory as the segment match file and Double Match Triangulator will add it to the match file information.

David: Paula's correct. I actually recommend you increase it right to 10,000, which is the maximum that GEDmatch allows. If B is still not in A's list with 10,000 matches, then use the One-to-one Report to get the match information you need and Double Match Triangulator will use it with the other matches. For information on how to do this, see the help page for using DMT with GEDmatch, which you can also find here: www.doublematchtriangulator.com/help/module_2_4.htm

Love your answer to bungee jumping.

About 50 years ago, I collected about 100 postcards of my city done by a local photographer. I'm hoping over the next few years to go to each site and retake many of them from the exact same vantage point. There are apps such as Then And Now for Android that can make a montage of the two.

I also have digitized black and white super-8 film that my uncle took 70 years ago driving through the city. I may screen capture some of the most interesting locations and do the same with them.

Blaine, do you remember what your topic was?

Hi Blaine, What was the very first genealogical/DNA talk that you ever gave? Where and when did you give it, and for what event?

Paul - Yes. Northeast Romania and Ukraine

Debbie - That's the classification given by Ancestry. The other companies all use Ashkenazi.

Paul Conroy - Various bits of nothing that makes any sense. For more detail, see: http://www.beholdgenealogy.com/blog/?p=2516

My genealogy says I'm 100% Ashkenazi. This is what each of the companies say: Ancestry 100% 23andMe 98.9% Family Tree DNA 92% MyHeritage 83.8%

Truth of the matter is that it doesn't really matter if the match is from the common ancestor or not. By just getting you to check for connections and finding one, then the DNA match has served you well. The main takeaway is that once you've found that first connection, don't stop there. Keep looking and you might find a second or a third.

Miss Boomer's concerned about you. She hasn't seen you for a few days and will keep checking every day until you show up.

Graham Hart - Can you give a teaser?

Randy's definitely THE inspiration for all geneabloggers.

A very quick look through this for my great uncle with a very common name in Jacksonville in 1935 is already giving me some valuable new information that may solve a few mysteries.

Bill Guest! He was one of Winnipeg's greatest news reporters. I watched the eclipse from in front of UMSU at the U of M in between classes and took pictures with my 35 mm Olympus OM 1 camera and telephoto lens.

And why is the sleigh named Bob?

I knew I messed up something six months ago when I made that dentist appointment for 10 a.m. But I should be home by noon CST which is 11 am MST so I'll at least enjoy your 2nd hour.

Double Match Triangulator (DMT) makes use of a testers' segment matches. 23andMe only allows you to download your own segment match file. But DNAGedcom can produce segment match files for you and for any of your matches at 23andMe. DMT can import the match files that DNAGedcom produces.

So what's the chance that my wife's Zaslavsky ancestors may have at one time lived in Zaslav? In the 1850's they were in Tetiyev.

I hope you've left someone in charge of feeding them while you're away.

Has Snow ever seen snow?

118 for me. 188 for my uncle who's on the same line as me. We are part of a Ashkenazi Levite study so the active recruitment of testers for that study might be why we have a lot of matches. Interestingly my uncle and I are GD 2 because of a couple of mutations I have that he doesn't. After my uncle my closest are GD 6. After me, my uncle's closest are GD 4 who are the same people as my GD 6 people. We don't have a single match who has the same surname, due to Ashkenazi's in Eastern Europe only adopted their surnames in the early 1800's, about 5 generations ago.

I think determining breeds are more accurate than ethnicities. Dog breeds are genetically different from each other.

I'm coming late because today is almost over where you are, Jonny. But how about telling us how you got interested in genealogy and then about how you decided to try DNA.

Here's an article by Debbie Kennett - https://cruwys.blogspot.com/2020/02/30x-whole-genome-sequencing-from-nebula.html

More info and commentary on Nebula Genomics by Roberta Estes: https://dna-explained.com/2020/02/18/news-nebula-genomics-whole-genome-myheritage-photos-go-viral-upcoming-publication-schedule/

Cold is beautiful though. -21C not much wind. Bright sun in perfect clear blue sky. Standing on the middle of the lake in Lindenwoods.

That's a warm winter day where I'm from. Try -30 C. (before windchill)

Love it! Wish they had customized cards here.

Your neighbors are going to start wondering why you're snooping around the neighborhood. Maybe you should put up "Lost Koala" posters to help find the ones that haven't been around for awhile.

And what ever happened to the two who were rescued. Was it Mickey and Little Mickey? Were they released?

Sara - Now if I can just get some of MyHeritage's Theories of Family Relativity to appear, then I'll be set!

Thanks Katherine. I followed your suggestion and contacted AncestryDNA via their support about why I had ThruLines and then they vanished about 4 to 5 months ago. It turns out that "potentially due to a recent update", some of my relationships with my parents and grandparents are "Unknown" rather than "Biological". Once I changed them back to Biological, all my ThruLines that I had before immediately reappeared. So now I'm as happy as a clam. I suggested they add this information into their ThruLines help so that others will be aware that you must have Biological relationships for the ThruLines to work.

Ahh, I found it: https://journals.plos.org/plosgenetics/article?id=10.1371%2Fjournal.pgen.1004144

Blaine - What is that paper you often referred to saying that WGS will only give 10% improvement in matching? WGS results have significant error rates in reading. Alignment isn't perfect. Even the human genome reference is variable. It would be desireable if it could be done, but it would be far from easy to do it.

Debbie - Yes. Using a variant to identify a match would be desireable. But that means it would also identify mismatches. Your cousins who were not passed that variant would not then appear as matches as they should. And if you had a mutation, insertion or deletion that your father did not, then your variant would be seen as a difference breaking your match with your father and all your paternal relatives on that segment. You also technically don't know which parent's chromosome that variant is on, so allowing individual variants to break matches will really hurt matching.

Debbie - Yes. The average person has 3 million variants from among those 45 million SNPs. So comparing two people, they may have close to 6 million differences. Many of those are rare SNPs occurring in one in a hundred or one in a thousand people. It takes many thousands of people to identify the 45 million SNPs that may differ among us humans.

David, Debbie. Well, the book analogy is not quite correct. On average, we each have about 3 million SNPs that are different from the human reference genome. Up to 700,000 of them are already tested via an array test. That means about 2.5 million variants are left in 3 billion base pairs, or about 1 in a thousand. Yes, looking at the first letter of a paragraph is a good analogy, except that all the letters in the remainder of the paragraph are identical with someone else except in 2 out of a thousand positions. Testing companies usually allow for 1 difference every hundred or two hundred SNPs because reading isn't perfect and there is the occasional variant, insertion or deletion. Therefore 1 in 500 appears as noise in matching and WGS will provide little help for matching over traditional array tests. Blaine has referred a number of times to a paper that concludes that WGS may only provide a 10% improvement in matching over array tests. So all the extra effort required to use WGS for matching really is not worth it.

How did Geoff get that Like in within 20 seconds of my post when I didn't even link him? He even beat your super-prompt response by 30 seconds. I guess "that's what friends are for".

I remember watching Kirk Douglas 2 years ago on the Golden Globes with his daughter-in-law. https://www.facebook.com/GoldenGlobes/videos/1884629181579091/

"Getting old is a gift. I sometimes forget that but it's true." Grandpa (Danny DeVito) in Jumanji, the Next Level

Vik - I sponsored you a while ago for the translation of some of the Mezhyrich records. Unfortunately, none of the people I expected to find were in there. I would be willing to help you out again if you have any towns in Wolyn nearby Mezhyrich that include people from these surnames that I might be interested in: Herman/Gherman, Lapedes, Zew/Zeff, Pisetzki, Mindes, Gershfield, Zaidman, Zimberg, Sitner.

They all look directly at you now, obviously having said: "Let's come to the human's back yard to see what she's up to". "Oh good. She's okay." I bet they get worried when you don't show up.

Kalani - I see you answered you have 100 - 150 matches with Chinese/ Vietnamese ethnicity which is more than anyone else. You obviously have relatives of your ancestors who married Chinese/Vietnamese.

Is he coming right up to you now?

Blaine and Debbie: Unfortunately the population geneticists aren't the ones advancing genetic genealogy, and the genealogists that have been for the past 10 years like Jim Bartlett and Tim Janzen aren't scientists who write formal papers that get reviewed by population geneticists. Phasing has always been known to be an effective way to filter out many of the false segments. But creating and verifying triangulation groups is just as effective because it is essentially the same as phasing. The relationship between triangulation and phasing is so simple to logic through. If you have a common ancestor who passed an IBD segment to both you and a known cousin of yours, then that segment came to you through one of your parents, most always the parent you expect. That match is now phased because both you and your cousin have different SNPs on your other parents' chromosome on that segment. Everyone in the triangulation group that matches both of you and each other must be matching on that segment of your parent and that common ancestor, other than very small segments under 6 cM that may have both parents randomly matching the one.

Jim is simply saying that while many small segments from 6 cM to 15 cM are false, some are valid and are IBD and not necessarily outside the genealogical time frame. Jim basically doesn't want to throw the baby out with the bathwater. Jim's techniques include triangulation and forming triangulation groups. Most people don't realize that by confirming triangulation of all members (4 or more) of a triangulation group, you are effectively phasing all the matches in the group to one parent. Almost all matches in Jim's groups will match one or the other of his parents. And that's whether he had his parents tested or not. Almost all will pass the parental filtering test. That also means that a lot of non-IBD matches have been eliminated from the 6 cM to 15 cM matches that Jim is analyzing, including all those that would be eliminated through phasing and/or parental filtering. In other words, creating and verifying triangulation groups with segments of 6 cM or more is as good as phasing and is equivalent to phasing.

There might be a sneaky thing you can do (but I haven't fully explored because I can't use VP for myself). Download segment match files for each of the siblings. Compare the segments each share with a third person. If one of the siblings matches the same people only up to a certain address, and the other sibling continues on, then that address is a crossover for the first sibling. If the two siblings match each other on both sides of the crossover, then it indicates that they fully match each other before the crossover and half match after it.

And here's his clusters at GEDmatch using 50 cM to 500 cM. Way more background grey, plus that huge cluster in the middle.

Okay, well look at this. At GEDmatch, I don't have enough matches over 90 cM to use their clustering at that level. But my uncle (also 100% Ashkenazi) does. At GEDmatch, I think they get more baseload matching than Ancestry does, maybe up to the 90 cM or 100 cM level that you mentioned previously. So here is my uncle's clusters at GEDmatch using 90 cM to 500 cM.

Jonathan, Adina: This is my theory: My parents are not related. According to GEDmatch's test for that, I'm only homozygous over a single small 7 cM segment. My uncle's parents (i.e. my paternal grandparents) are also not related. My uncle has no homozygous segments over 7 cM. I think this might be the main reason why I can separate out grandparents through the Leeds Method or other techniques despite my endogamy. I do Leeds down to my 50 cM DNA relatives and that works well and below that it starts to mumbo jumbo, so that's why my theory (at least in my case) is that my endogamy includes up to 50 cM of baseloaded matching to each of the people I match to.

Adina - My 4 grandparents come from 4 towns now in Romania and Ukraine not more than 200 miles from each other. I'm using Leeds exactly as specified.

Adina - Jonathan was calling it "hopeless". Above I was describing why for me it was successful.

Jonathan - Only 9 of my 50 DNA relatives appear in two or more columns. I'd call that successful because I can assign very likely grandparents to the others. I have 8 known relatives at Ancestry DNA covering all my 4 grandparents and each is in just one column and those with the same grandparent are in the same column. That pretty well nails it for me.

… and I would NOT say that "the Leeds Method is hopeless" for Ashkenazi endogamy. I have obtained reasonable results with it and it can segregate most of my closer matches at Ancestry into my 4 grandparents. See: http://www.beholdgenealogy.com/blog/?p=2858

Jonathan: It puts me in the "significant endogamy" category with an estimate of 14%. This is a zoom-out of what my clusters look like:

Hi Maureen. My daughter was born when her mother was 30, her mother's mother was 60, her mother's mother's mother would have been 90 her mmm's mother would have been 120. Genealogy numbers are fun.

Thomas - I almost thought you were using a DNA test as a reason to go drinking.

Gabrielle - You'll be able to go back more generations than Jim or I will.

No, no, no. They've got to get along. You'll have to post backyard rules: "No acting like humans"

My great-grandfather was born in 1852, which is 104 years before me. What that does is make it more difficult for us to go back more generations.

Cyndi still has a few days left to enjoy the January picture of Baby Bear.

Thanks Celia. Based on our discussion, I've now updated my Life and Death of a Segment article correcting the Wikipedia equation I used. It did not change any of my observations or conclusion. http://www.beholdgenealogy.com/blog/?p=3057

Celia - That's a very interesting line of thinking to come up with a solution. Hmmm. The Wikipedia formula isn't sourced either, and I don't like that it approaches 50% as cM goes out to infinity. So Wikipedia could be wrong here with respect to what we are talking about. I think the correct analysis is this: Assuming a Poisson distribution for crossovers (which is what is usually assumed), then the Prob(zero crossovers) when the mean is cM/100 is: exp(-cM/100), therefore: The probability of one or more crossovers = 1 - exp(-cM/100) For 100 cM, Poisson gives 63.2% compared to your 63.4%. Putting all the equations in a spreadsheet and comparing them, I see that your equation and the Poisson are so close that the graph only shows one line for the two. You can start to see a tiny bit in the expanded detail. I think the Poisson version I give is technically correct. But your equation does just fine as well.

Celia - I thought the correct equation is Prob(recombination) = (1 - exp(-2*cM/100)) / 2 as given on the Wikipedia page for Centimorgan. For 30 cM, that gives 22.6%. I've never seen the equation you give before. Can you give me a reference for that? Most people don't realize there is a mathematical difference and that cM actually are the average number of recombinations expected in one generation over a segment. For small values, the cM/100 is a reasonable approximation. Using 22.6% or 26.0% instead of 30% does not change any of my arguments above.

Adam - and you're forgetting that we all only each have about 250 DNA ancestors. I'm sorry. Not 250. Maybe a few thousand. But certainly not a million. https://gcbias.org/2013/11/11/how-does-your-number-of-genetic-ancestors-grow-back-over-time/

Adam - Well one in a billion divided by 230 divided by 1 million means there is a 1/4 chance that one 30 cM segment will match somewhere in the genome from one of those 1 million ancestors.

Jonathan - I was giving an alternative to your last paragraph where you were talking about sites that provide full segment info.

Jonathan - ... or only include segments that form triangulation groups with two or more people.

Forget what MyHeritage and everyone else has said. With 4 siblings, you can easily do Visual Phasing which will assign the segments of all siblings to your four grandparents. Then it's just a matter of figuring out which grandparent is which. Then you'll suddenly be able to associate matches to grandparents, and those on your mother's side you'll be able to research further. https://thegeneticgenealogist.com/2016/11/21/visual-phasing-an-example-part-1-of-5/

Jonathan - I see now that you and I exchanged comments last August in my Life and Death of a Segment post. Frankly, I had even forgotten that I wrote that post. And you must have forgotten that you've seen and commented on it as well. http://www.beholdgenealogy.com/blog/?p=3057

No never.

Blaine - I don't believe that. A 30 cM segment has a 30% chance of dividing, a 35% chance of not being selected and a 35% chance of being passed down. For 20 generations, that's 0.35 ** 20 which is less than 1 in a billion chance of it surviving to a specific 20th generation descendant.

Alona should open a bed and breakfast for them. All she needs are the beds.

Jonathan - That is true. I should try everything again for FTDNA and for 23andMe, when I get some time (Ha, ha, ha!)

Jonathan - Actually I see now that about a year ago I did look at your tool on my Ancestry DNA data and compared it to Genetic Affairs and Collins' Leeds. http://www.beholdgenealogy.com/blog/?p=2863

Jonathan - I'm going to try your Shared Clustering Tool. I'll let you know what I come up with.

Jonathan - But now I suddenly recall that you are one of the experts in clustering. Clustering works, even on your own FTDNA data, does it not?

Jonathan - If your ICW lists are the same, it sounds more like related parents than endogamy. Endogamy happens at the distant relationships and should not prevent you from clustering. But if your parents are 2nd or 3rd cousins once or twice, that could wreak havoc. My parents are only very distantly related (one segment, 7 cM) according to GEDmatch's tool.

I still have zero. I've linked mine, but 23andMe have not advertised that feature very well. Too many people don't know about it.

I have 264 new matches in 2020 so far. That's an average of 11 a day.

In general, I find FTDNA has about 50 cM of background noise for each of my matches. So for a rough estimate, I simply subtract 50 and consider that. But there's still stuff we can do. The In Common With and Not In Common With are useful for clustering and works for endogamy, because the dominant grandparent still tends to define our relatives' clusters. And we can download our segment match files and find triangulation groups to further group people. It's true that not everyone has trees, but a lot enter their ancestral surnames and places and that is very helpful. If our Eastern European records went back further than 5 generations and our ancestors adopted surnames earlier than 5 generations ago, then we would be able to connect to a few people. But the majority of our matches are obviously 5th cousins or further and as a result mostly un-connectable.

Right! I've got over 22 thousand matches at FTDNA and the only ones I know are my uncle who I tested, and one 3rd cousin.

You'll have 24 hours in airports and airplanes on your way to Salt Lake City. That could help you chew through a book or two.

Franklin should work on his family genealogy. That will keep him from getting bored.

Oh, and their time zone is a half hour before Atlantic time. Many years ago, the great comedy team of Wayne & Shuster used to always joke: Coming up at 8 p.m. across Canada, 8:30 in Newfoundland.

Well that's Newfoundland. To live there, you have to be really gutsy. They have the most rainfall, most snowfall, most days of fog, and highest average windspeed of anywhere in Canada. But they get nowhere near as cold as we do in Winnipeg.

Em - This is a technique that anyone can use on DNA Painter with their segment matches and the DNA Painter that exists now. It is something I discovered possible while developing my presentation, and I will reveal what it is during my talk at the Family History Fanatics Winter DNA eConference on Saturday.

It's a trick you can use on DNA Painter.

No. I’ll be talking about Double Matching and Triangulation Groups, but I’ll be presenting it in a very visual style taking you through some of the challenges I have with my own genealogy and DNA matches. I’ll be introducing some concepts that I don’t think have been discussed before. If you want to know details on how to use DMT, especially acquiring data files, I think I have done a good job with that in the DMT help file.

Times shown are Central Standard Time. To convert, go to Google and type, e.g.: 8:30 am CST in (your time zone), e.g. in GMT.

Em Candy - This is what I meant by the use of the word "exclusive". But now I see I used it as an adjective instead of a noun. My apologies.

Dianne - By registering, you'll be sent a link to the recordings of the talks and you'll have a month to watch or re-watch them.

Em Candy - Because it's not a new feature.

Em Candy - It's exclusive because it's not going to be revealed until my talk. (I'm just trying to build a little excitement)

Sharyn - It's online. So "being there" is at home in your jammies. Whether you can or can't watch live, by registering, you'll be sent a link to the recordings of the talks and you'll have a month to watch or re-watch them.

Interesting. I had uploaded in 2018. They closed in Sept 2019 and at that time said all their data would be erased. I wonder who took them over?

Paddy - Yes, exactly!

p.s. I'll be talking a bit about this on Saturday at Family History Fanatics Winter DNA eConference on Saturday. Jonny Perl, Paul Woodbury and Devon Noel Lee will be speaking as well. https://www.familyhistoryfanatics.com/winterdna

Paddy - Good. Checking triangulation is the first step. That still doesn't mean they come from the same ancestor.

One possibility is that A's father's chromosome matches B's father. A's mother matches C's mother and B's mother matches C's father. That's still a triangulation but there are three different ancestors. To prevent that case, you need at least one more match in the triangulation group and a check that every two of the four match each other.

The other possibility in small segment matches is that one of the people randomly matches with both parents' chromosomes the one chromosome match of the other two (or even all the other people in the triangulation group). This case is a horrible one because just about all tests will seem to indicate they match.

All I'm saying is not to conclude that very unexpected small triangulations indicate one common ancestor. If that was happening at 15 cM or even 12 cM, then I'd say it's likely true, but not at under 8 cM.

Sounds wonderful. I'd like the same for Romanian, Ukrainian and Russian.

Well done. Just be careful not to make conclusions about your matches in the 7 cM range. If you haven't confirmed that those small segments that overlap actually match each other, then there would be about a 50% chance that any of them might be proven false if you were able to phase both parents. See the False Positive Matches section of the Identical by Descent page of the ISOGG wiki. https://isogg.org/wiki/Identical_by_descent#False_positive_matches

Not this year, unfortunately.

David - You have a personal lifetime license good for any computer you use.

You make it sound like DMT caused your computer to crash. 😧

David - You'll be able to chat and ask questions which the presenters can answer at the end of their presentations. I haven't heard if there will be any free draws.

David: Following the live presentations, those who register will be sent a link to the recordings which will be available for one month to see what they missed or re-watch it.

Alona. We also have the world's largest snow maze. https://www.instagram.com/p/BsTrIm9B44V/

Thanks, Alona. I haven't seen snow patterns like that before. That's pretty amazing. But we do get beautiful snow sculptures built throughout our city every year for the Festival du Voyageur competition each February.

Ah! University dreams. Me too. So common.

Here's similar story but true. I had a very absent minded Math prof who took his car to University one day and took the bus home. When he got home, he looked in his garage and saw his car was missing. He phoned the police and reported it stolen.

I can only imagine what his University dreams are like.

The new generation is used to seeing you since they were born. Have you been slowly moving the food a bit closer to you?

Jill, remember to allow a bit more time for going through metal detectors on your next flights. Glad you're so upbeat and nonchalant about this. Best way to be.

How many babies, both Koalas and Roos do you now have birthdates for?

Hmm. I like the koalas and roos outside more than I like possums in the cellar.

We very much like Celebrity. We've done them twice and are doing them again on the Celebrity Reflection from March 2 to 13 to the ABC islands.

Enjoy. Unfortunately we live in Winnipeg.

I'm sure the assembly instruction book is 4000 pages long.

238! Looked through the list hoping that one would be one of the many Australian genealogists that I have met, but no such luck.

You should set up a couch, a big screen TV and a refrigerator for them.

Your blog looks and works just fine on mobile. I wouldn't worry about it.

Blaine - I purchased both FASTQ -> hg38 BAM and FASTQ -> hg19 BAM and I've now got my results back. I also was successful creating a BAM myself. See my blog post about it: http://www.beholdgenealogy.com/blog/?p=3209

Monique - PAF was an extremely popular program and many people continue to use it today despite it having been discontinued for so long. Or those that don't use it any more still remember it as to them being better than the program they use today. It has consistently been highly rated for the past 11 years. Read some of the reviews and you'll see some of the reasons people liked it. It was a very capable program, and was available to anyone at no cost.

Keep on doing what you love, Judy. Never stop. Never surrender.

That is horrible. A year ago, the west coast USA forest fires destroyed the town of 26,000 ironically named Paradise, California. But there's always hope and resiliency, both before and after. https://www.nbcnews.com/news/us-news/paradise-regained-year-after-camp-fire-resilient-town-rebuilds-n1077991

I also keep my Epson DS-860 scanner just to the left of me, in front of my desktop computer against the wall. Behind my right screen easily accessible but out of sight are hole punch, scizzors, magic tape, stapler, staple remover and paper clips. Normally I would keep them in a drawer, but this way there's less interruption when the family comes in to borrow them. Also, get the biggest desk you can. Mine is 3 feet by 5 feet 6 inches.

Let me add the magnificent work Katherine sent me that arrived today.

I must say though, that I am quite worried about your dogs Isabella and James. I do hope Isabella will get to college and James will enjoy Europe and then finish secondary school.

Alona - Do kangaroos drool?

Incidentally, compared to humans with 46 chromosomes (23 pairs), red kangaroos have 20 (10 pairs), grey kangaroos (and koalas) have 16 (8 pairs) but dogs have 78 (39 pairs). Cats have 38 (19 pairs). Bananas have 22 (11 pairs). And I was very surprised by the variation in this after I looked it up.

Alona - I keep trying to take photos of the squirrels (and rabbits) in my yard, but they're way too jumpy and scurry away at the slightest sound. Ducks and Canada geese are much easier, but they fly south for the winter. I can try bribing them with food, but I really don't want 100 squirrels in my yard.

Maybe you should do DNA testing to find the family relationships. I wonder if Embark would accept kangaroo samples if you asked them to try. Would be interesting to know how many of the kids are Charlie's.

I've enjoyed staying at the Hampton Inn for RootsTech twice and the Crystal Inn Hotel once. Both are about 3 blocks South of the Ice Palace. Walkable in 15 minutes (even with a walking boot on). Free breakfasts!

This article talks about some of MyHeritage and Living DNA's looking into artifact testing and results. https://www.theatlantic.com/science/archive/2019/03/dna-tests-for-envelopes-have-a-price/583636/

Hana - I don't think DNA Memorial does artifact testing.

If Mickey and Minnie are released within 5 km of your home, I'm sure they'll find their way back to your yard.

p.s. This arrived today. Thanks so much.

In my newspaper today.

Do koalas have to be cared for once saved, or can they be released again? I know you'll be able to tell which of their koalas are Mickey and Minnie, but if they can be released again, do they have them identified in any way so they can be dropped off about where they were picked up?

They don't seem to be too worried about the fires. That's a good sign.

We're all hoping all your hard working fire fighters can manage to keep the burning under control until the weather gives you some assistance.

Wishing for a two days of rain miracle for you and all of Australia.

I uploaded my 100% Ashkenazi DNA to Living DNA a couple of years ago. I now have about 300 matches there. But I don't know how I'm related to any of them.

Winter in Winnipeg is definitely 2 weeks shorter ever since 7 years ago when my wife and I started taking a 2 week winter vacation somewhere warm.

Shauna - Me too. I doubt we could get that much for it as a collectible.

Jenny: Here's an article by a genealogist who I deeply respect: https://www.legalgenealogist.com/2013/06/30/dna-life-after-death/

So then if the fire in the early morning made it up to the extent line and burns out or heads back at that point, then most of the grass would be burnt and I'd at least hope that it shouldn't be able to approach from that direction again.

Anthea is safe because Alan is protecting her.

Egads! It's the size of Adelaide.

Yikes! Alan: There don't appear to be many trees up to the extent of the file. Is the fire travelling primarily from tree to tree or is there enough dry grass that it is travelling along the grass. And there does appear to be a house just below that yellow line. Did it burn?

Well, obviously none of us want your koalas to be in extreme distress, Alona. But this might be your chance to let them come to you if you sit there with a bowl of water.

Aleph, Lamed, Vav, Nun, Hey: אלונה
It's pronounced the same, but the 5 Hebrew letters are written right to left.

As far as your "real name" goes, well, you name all your koalas and roos. What would they name you? Food lady. Photo lady. Keeper of the back yard. The watcher. Overseer of the great outdoors.

You're the first Alona I've ever met. So now you made me curious to look your name up. I'm surprised to see (via Wikipedia) that Alona is a Hebrew name, still popular in Israel:

Alona Barkat (born 1969), Israeli businesswoman and football team owner
Alona Frankel (born 1937), Polish-Israeli author and illustrator of children's books
Alona Kimhi (born 1966), Israeli author and actress
Alona Koshevatskiy (born 1997), Israeli rhythmic gymnast
Alona Tal (born 1983), Israeli television actress

Alona is the feminine version of the Hebrew name Alon which means "tree" or "oak tree" in the Hebrew language. In Israel, Alon is used both as a surname and as a masculine given name.

Who were you named after? Why did your parents name you Alona?

Nice that you had wonderful clear skies during the entire Antarctic part of your journey.

If she has a joey, you can use the name Figgy.

You'll have to try to attract those other three to your back yard. What do they eat?

A tree frog on a door. Maybe Dory is a good name.

Roger - Krill s--- won't smell when it's frozen.

Gary, Maybe because that picture was taken 10 years ago. :-)

Almost surreal. And you're there during their summer. Imagine what it's like during their winter.

Your captions add so much drama. Days of our Koala's Lives.

Okay. That's good. My 8 pair of FASTQ files from my short read test totals 149.65 GBases. But your program doesn't work for my long read test. It just sits there at zero. Each reading and quality line in my long read file are not all 100 bp like in a short read file, but are as long as each individual read is. So some lines could be hundreds of thousands of characters long.

Thank you. Now. it's working.

Nope. Same problem.

I got this error for all my files.

Laurie:

Since once 1/10 of an iceberg is above water, and since it's 65 meters high, there must be 600 meters of it below the surface!!

How close did the penguins let Robert get? He looks like he must be pretty close when taking that foot picture.

Aren't two of each allowed on board?

In Winnipeg, we walk on ice all winter. And we have a skating path on the river.

Never heard the term "growlers" before. You can't be referring to polar bears, because they're only in the Arctic, not the Antarctic?

I'm very interested in going. I'd like to talk if possible. Have you seen a Call For Proposals?

Yes. I first met Curtis at RootsTech 2017 where he and I were boot buddies. He's a wonderful person. I'm saddened by the troubles he had to face at GEDmatch the past year or so. I hope he and John were dutifully rewarded with a rumored 8 digit payment for GEDmatch.

Don't know if this has already been said in the previous 787 comments (wow!), but Verogen has started a new Facebook page for GEDmatch: https://www.facebook.com/officialGEDmatch/

I went to the mailbox and the sunset was so nice I walked another block to the park. -22 here does look a lot like Antarctica.

Hey, it's -22 C in Winnipeg right now (-28 C with windchill). Full sunshine though! But the lake is frozen solid.

Set up shop there, and you'd be able to take pictures of the seals and penguins coming to your back yard each day.

I haven't seen Calvin and Hobbes comics in years. Loved them.

All the best to you, Jan, on your birthday!

Wow!!

I was able to register as a new customer with YSEQ and purchase a FASTQ mapping to hg38 for $25. I didn't have to pay any extra.

Blaine - Now you've sparked my curiousity and I think I'll order hg38 as well and compare them.

Blaine - So did they gave you an hg19 SNP file with your hg38 BAM and VCF files?

Blaine - I didn't realize that GEDmatch accepts hg38. Now that I look, they specifically state at the end of their upload page that "Your data must be Build 37 (GRCh37 hg19)". So what happened when you uploaded your hg38 file? Did GEDmatch accept it? If so, how does it compare with one of your other tests of yourself that was hg19? Also, I didn't, realize that YSEQ gave you that file to upload for hg38. One of the differences between their descriptions of their FASTQ mappings for hg19 and hg38 is that specific file.

Not all of them give you your raw data (FASTQ or BAM files) which is important. YSEQ does do Whole Genome 30x for $1340.

Blaine: My hard drive is already filling up. 🙂

Thank you for telling us this Blaine. I hadn't considered YSEQ for any of my WGS analysis before. I see they have both FASTQ Mapping to both hg38 and to hg19. You chose hg38 which is what YSEQ uses and what FTDNA uses for its Big-Y 700. And hg38 is definitely wanted for health analysis. But as you know, all atDNA matching for genealogical purposes is compared via hg19. And YSEQ's FASTQ Mapping to hg19 says for that they will also "create a file format that can be used for consumer genetic online services like GEDmatch" which they don't do with the mapping to hg38. I'm intrigued to see what they produce for that. So I'll likely try the mapping to hg19.

You're so practised at giving excellent lectures, Judy, that I hope you can take a break and not find that you're still giving the lectures in your sleep.

Enjoy your coldest summer ever.

You are the scribe of your tribe of Kangaroos. Because of you Alona, so many of us are enjoying the adventures of their lives, just as we do with our ancestors. You'll have to copy this lovely obituary to Zeb's entry in your Kangaroo family tree at Ancestry. (You do have one, don't you?)

Happy birthday, Vik

Excellent question, and I'm interested in seeing other people's answers as well. What I do is I have a page in One Note, which includes the people I match to that I know how I'm related to. I include my common ancestor and other people that cluster with them through one of many different Leeds-like clustering tools. The goal being to identify how the others might be related through this common ancestor.

What on earth are those?

As long as the cat doesn't get jealous.

Mags - If Gale was 999, then I had to be 1000. ... Unless Gale saw me come in as 999.

Gale French 1000!

Louise - About 50 instant (i.e. half strength - 45 caf, 5 decaf), 5 brewed Tim Hortons (full strength caf), and 5 brewed Starbucks (double strength caf).

Blaine, I thought MyHeritage was sufficiently hiding my living people from others. Then a few months ago, I received an email from a third cousin (also a genealogist) who asked me to remove his birthday from my site. It turns out a smart match made this available to him. I then went to Family Tree Builder and marked each living person as "Private" and resynced my tree. My underlying reason is that my wife and daughters are not into genealogy and don't want me to include their personal information online. I try to respect that and feel that it is right to extend that respect to everybody. And to top it off, MyHeritage specifically tells you that you cannot put up information about living people without their permission. Since I have no permission from anyone, I feel it is right to respect that rule.

I've recently come to my own personal conclusion that it's important to hide all information about living people. I keep my master information on my computer but I won't transfer information about living people online. For obituaries, I'll keep copies of the originals in my personal files, and what I put online will have names and info of living people blurred or redacted. I know there's lots of info out there about living people. But I don't want to be the source with that info.

I got a cold call a number of years back. "Hi. I'm with Investors Group. Do you have a few minutes to talk about investments?"

I said "Sure. What would you like to know?"

That threw him off.

Hi all. I've finished getting everything from our Yahoo Groups Major Work forum before they delete it all on Dec 14, and I've put it all up on my OneDrive account for everyone. A few minutes ago, I sent everyone an email with the link because I don't want to post it publicly.

I used an email list I had from 2017, so if you don't get the email from me in the next few hours, let me know.

Gail, Gary, Shell, Bev, Brian, Merril, Carol, Kathy, Margret, Brenda, Diane

An odd thought: How about putting an electronic bathroom scale beside the birdbath and maybe you can find out how big these big boys really are .

The records of Romania/Ukraine availalble to me only go back to about 1850 so 5 or 6 generations.is my limit.

Wow. You've got 170 Common Ancestor Hints out of 1893 = 9%. I've got 16,982 4C or closer and only 9 Common Ancestor Hints = 0.05%

Right! I forgot that you and I share a birthday. Happy birthday, Jennifer.

Bombers upset Hamilton to win the Grey Cup 33 - 12.

Kalani - I don't know. Maybe your parents are 4th or 5th cousins in some way you don't know of. Or maybe those are by chance matches since they're all less than 15 cM. Do those regions triangulate with people who form two groups that don't triangulate with each other? Or does everyone triangulate with each other? That will tell you if it's a false match between your parental chromosomes, or if they are actually the same segment from both parents.

I used my uncle instead of myself because I had one segment of 8.8 cM with the conclusion that "your parents are probably distantly related". So his was a better example than mine.

Kalani: Yes. 0 cM. There are no segments larger than 7.0 cM. The conclusion the tool writes is "No indication that your parents are related."

It's important to realize that endogamy and parents being related are two different things. Here's my uncle 100% Ashkenazi, with father's family and mother's family living within 200 miles of each other in Romania.

From what I understand, even long reads are not long enough to completely phase, and that's why you can't find software to do it. There are very long sections that can be hundreds of thousands of base pairs where both parents' chromosomes have the same value and long reads can't span that. The best you can be expected to get with long reads is about 200 contigs over the 23 chromosomes.

I third the motion.

"work" ==> have fun doing what I do

Boomer and friends now coming to Canada, too.

From what I've seen, 30x WGS typically averages 35x to 40x over the autosomes, but tends to be less on the Y averaging 20x. I think that's just the way it works because you've only got 1 Y chromosome, not two. You'll likely find your X is about 20x as well.

Maybe you can put out a kiddie's pool full of water for them.

I find one single segment matching that is more than 200 cM to be remarkable. And there's 6 cases of that.

Probably 23andMe's new Family Tree beta. You have to have tested there to have access to it because it uses your DNA matches there. They don't accept uploads from other companies.

My largest single segment match at Ancestry is 39 cM, at MyHeritage is 22 cM. At 23andMe, they limit matches to the people sharing among your 2000 closest. I don't have anyone in my list sharing fewer than 3 segments.

Definitely looks like more than 30x

You actually mow that grass?? Maybe you need a few sheep.

I bet there's no sales clerks and you have to order online.

https://support.ancestry.com/s/article/DNA-Genetic-Communities

Blaine - I'm thinking that the DNA-based ethnicity percentage represents the first line. For me European Jewish = 100% and for Kevin Norway = 18%. I'm wondering if the 2nd and 3rd lines are not DNA-based, but are family-tree based on the locations I include for my ancestors in my Ancestry tree. Otherwise, I don't know how they can get that accurate for Kevin and me on our 3rd lines.

Blaine. Okay. But my 100% and Kevin's 18% are on the first line, not the second or third line.

Is it possible that they are using our family tree location data to "help" them? To me, that last line is so correct, that it is suspicious.

Katherine - My availability will be limited at least until next Fall. But my topics are at www.lkessler.com/speaker.shtml

Jonny - Interesting. I only have 10 DNA matches when I filter by Smart Match. In all cases, the Smart Match is a marriage into my family.

Almost all my Instant Discoveries turn out to be families of the spouse of one of my relatives. I almost always do not want to add them to add them to my tree. So I reject them, which feels wrong because the discovery itself is correct.

Gary, Gail, Shell, Bev,

Okay. I've tried a few downloaders. They don't work. And even if they did, they'd still require lots of work to clean up the results.

I've figured out a fairly efficient way to go through the Major Work conversations and create a reasonably clean PDF file of them.

The best thing about going through the conversations myself and doing this is that I get to read them all again, and believe me, they are definitely worth saving and reading again. We have so many great stories and memories there, as well as our life profiles.

I'm going to do this for the rest of the conversations. I plan to spend a couple of hours a night and I expect it will take 2 to 3 weeks to complete. I'll definitely get them done before Dec 14 when Yahoo will delete them. I'll also download the pictures and I'll make them all available to everyone.

I've also got my long read FASTQ files and that is all. From what I understand, you should align long reads using alignment routines designed specifically for long reads. Short read alignment on long reads will not perform as well. See, for example: https://www.biostars.org/p/327002/