PDA

View Full Version : Zero Big- Y matches in DF27



lgmayka
11-30-2017, 08:06 PM
#399633 just received his Big Y results. He tested positive for BY22302, and was therefore placed into
DF27 > ZZ12_1 > BY3332 > BY19284
According to his Big Y Matches page.

But he actually tested negative for BY19284.

He has absolutely zero Big Y matches.

St. Pierre
11-30-2017, 08:34 PM
My haplogroup is R1b-U152-L2-BY3604 but my maternal grandfathers was R1b-DF27-S225 and his family name was Raymond in Canada/USA but was originally de Faugas from Langon in southern France.

razyn
11-30-2017, 09:28 PM
He has absolutely zero Big Y matches.

Those zeros are more apparent than real, they will get populated once the haplotree is updated following the huge rewrite of all the positions for Hg38. I had an email today suggesting that that process has been frozen for about eight weeks but is currently underway, and the backlog of known haplotree corrections will take another week or so to clear out. Then the BigY matches should get unlocked, and be more useful than the previous ones. That remains to be seen, but all it costs extra is some extra patience.

lgmayka
12-01-2017, 06:56 PM
#399633 just received his Big Y results. He tested positive for BY22302, and was therefore placed into
DF27 > ZZ12_1 > BY3332 > BY19284
According to his Big Y Matches page.

But he actually tested negative for BY19284.
In fact, BY22302 is the only SNP at that level for which #399633 tested positive. He had no call at BY22301, and tested negative at
A15900
BY19284
BY19285
BY19286
BY19287
BY19288
BY19289
BY19290
BY19291
BY19292
BY19293
BY19294
BY19295
BY19296
BY19297
BY19298
BY19299
BY19300
BY19301
BY19302
Z26495

ArmandoR1b
12-01-2017, 08:11 PM
So #399633 should have a lot of unnamed variants and the difference between him and other people that are positive for BY22302 is probably more than 40 (more than 20 each) which means that even after the hg38 update is completely finished #399633 still won't have any matches unless a new kit appears with shared variants that are currently unnamed.

lgmayka
12-01-2017, 08:54 PM
So #399633 should have a lot of unnamed variants and the difference between him and other people that are positive for BY22302 is probably more than 40 (more than 20 each) which means that even after the hg38 update is completely finished #399633 still won't have any matches unless a new kit appears with shared variants that are currently unnamed.
#399633 has 44 unnamed variants, and the actual limit of the Big Y Matches page is 30. His patrilineage is ethnically German, but geographically from Wyrzysk, Poland (https://en.wikipedia.org/wiki/Wyrzysk).

razyn
12-01-2017, 09:42 PM
My recollection of the matter is dimming a bit, but it was lgmayka who described a group of peculiar STR matches as the "pseudo North/South cluster" on DNA-Forums, about the time I started reading those discussions (spring of 2011). And my kit number was among that "pseudo" (I prefer to think of it as "elite") group... now aka FGC15733, not that the FTDNA haplotree has quite cottoned to it.

Except for known relatives on my father's side, my best match was then (on STRs), and continues to be now (on NGS-tested SNPs), an ethnic Kashubian family named Richert. They are also from Poland, and therefore familiar to the admin of that regional DNA project. In the R1b-DF27 project, those kits are 195834 and 44479.

ArmandoR1b
12-01-2017, 10:22 PM
#399633 has 44 unnamed variants, and the actual limit of the Big Y Matches page is 30.

I've seen where the actual cutoff has been discussed and I thought there were some people that weren't cutoff at 30. I can't remember which thread that was. Either way, 44 unnamed variants is too many to have matches due to a cutoff being added.

TigerMW
12-02-2017, 04:33 PM
I've seen where the actual cutoff has been discussed and I thought there were some people that weren't cutoff at 30. I can't remember which thread that was. Either way, 44 unnamed variants is too many to have matches due to a cutoff being added.

I agree but I think the base of the problem is getting a consistent measurement for counting. All variants and all variant calls are NOT created equal. There needs to be a winnowing to only very high confidence variants before we even discuss what the cut-off/threshold should be.

Setting a threshold for matching is similar challenge in the Y DNA STR matching. Imagine if that was riddled with no calls or questionable calls. There is such a fine line in the GD counting that would cause many, many problems. It's already only a rough tool in the first place and that is with very high confidence calls only.

ArmandoR1b
12-04-2017, 03:04 AM
I agree but I think the base of the problem is getting a consistent measurement for counting. All variants and all variant calls are NOT created equal. There needs to be a winnowing to only very high confidence variants before we even discuss what the cut-off/threshold should be.

Setting a threshold for matching is similar challenge in the Y DNA STR matching. Imagine if that was riddled with no calls or questionable calls. There is such a fine line in the GD counting that would cause many, many problems. It's already only a rough tool in the first place and that is with very high confidence calls only.

I wasn't stating that I agree with the threshold. I was stating that kit #399633 has more variants than what is allowed for matching by FTDNA. I don't doubt that there will be some people with just a few too many variants due to a problem with questionable calls and what would be matches with too many no-calls causing people to not get some matches that they should. However, a total of 30 non-matching variants between two people, about 15 positive variants for each person, means the two people have a most recent common ancestor more than 1,000 years ago in the direct paternal line. Is it really necessary to see full names, email addresses, and tree from matches that aren't related genealogically? I can understand wanting to see the information displayed in the FTDNA project results or what is shown in YFull but I don't think that the rest of the information is necessary. So, if there are a few extra questionable calls for one person and some no-calls for another and the two people aren't showing as matches because of it the MRCA is probably from over 1,000 years ago anyway. I won't lose sleep over it if it happens to me.

JMcB
12-04-2017, 03:39 AM
I've seen where the actual cutoff has been discussed and I thought there were some people that weren't cutoff at 30. I can't remember which thread that was. Either way, 44 unnamed variants is too many to have matches due to a cutoff being added.

The problem is FTDNA is saying their cutoff is 40, when - as Igmayka says - it appears to be 30 in practice. To compound matters, they are reporting a lot of junk as variants. At least, according to my Administrator and I believe McDonald, too. So until they clean that up, their matching is going to be skewed.

Edit:
http://www.anthrogenica.com/showthread.php?12230-Big-Y-changes&p=317028&viewfull=1#post317028

ArmandoR1b
12-04-2017, 05:56 PM
The problem is FTDNA is saying their cutoff is 40, when - as Igmayka says - it appears to be 30 in practice. To compound matters, they are reporting a lot of junk as variants. At least, according to my Administrator and I believe McDonald, too. So until they clean that up, their matching is going to be skewed.

Edit:
http://www.anthrogenica.com/showthread.php?12230-Big-Y-changes&p=317028&viewfull=1#post317028

However, even with 40 as the cutoff and zero junk variants kit #399633 still won't have BigY matches. The reason for that is each person that is positive for BY22302 has more than another 24 real variants that are phylogenetically equivalent to BY22302 and BY22301 plus at least one more private variant. Kit #399633 is positive for only one of those phylogenetic equivalents, a no-call for one, and negative for the rest in the 21 that Lgmayka listed. That means that kit #399633 has at least another 22 of his own real variants. He probably really has at least another 24 variants just like the kits in the Alex Wiliamson's tree here (http://www.ytree.net/DisplayTree.php?blockID=1513).

That means kit #399633 and any other person that is currently positive for BY22302 has at least 40 real non-matching variants. Kit #399633 needs to share about another 18 variants with another person in order for him to have 30 or less non-matching variants and have a match in BigY with a cutoff of 30 once the junk variants are fixed. For kit #399633 to have a match at a cutoff of 40 there needs to be a another person that he shares about another 8 variants.

If there was another person that kit #399633 has 40 or less non-matching variants with then kit #399633 would have more named variants and they would have their own branch that has BY22302 and other variants not currently found in the FTDNA haplotree.

JMcB
12-04-2017, 07:01 PM
However, even with 40 as the cutoff and zero junk variants kit #399633 still won't have BigY matches. The reason for that is each person that is positive for BY22302 has more than another 24 real variants that are phylogenetically equivalent to BY22302 and BY22301 plus at least one more private variant. Kit #399633 is positive for only one of those phylogenetic equivalents, a no-call for one, and negative for the rest in the 21 that Lgmayka listed. That means that kit #399633 has at least another 22 of his own real variants. He probably really has at least another 24 variants just like the kits in the Alex Wiliamson's tree here (http://www.ytree.net/DisplayTree.php?blockID=1513).

That means kit #399633 and any other person that is currently positive for BY22302 has at least 40 real non-matching variants. Kit #399633 needs to share about another 18 variants with another person in order for him to have 30 or less non-matching variants and have a match in BigY with a cutoff of 30 once the junk variants are fixed. For kit #399633 to have a match at a cutoff of 40 there needs to be a another person that he shares about another 8 variants.

If there was another person that kit #399633 has 40 or less non-matching variants with then kit #399633 would have more named variants and they would have their own branch that has BY22302 and other variants not currently found in the FTDNA haplotree.

I was addressing the confusion concerning FTDNA's cut off number (30 or 40) and the overall problem of unwarranted variants being called. As for the particular kit in question, your point is taken.