Ancient I-M253 samples list

[mwauthy]

Cheers - I try and follow the real data when I can - there's a fair bit of conjecture out there. Some is good, some not so good... Going back to the primary data such as BAM files is always worth doing if that's an available option.

As far as I'm aware, no one has yet broken up the I1 bottleneck yet. It's entirely possible that someone will at some point, but if they have and have tested, I'm not aware. There are a few individuals of rare lineages on the DF29- branches. There was some crowdfunding on the I1 Facebook page a couple of years ago for testing one of the individuals with a YElite that is now in the I-CTS12798 branch (DF29-, Z17954-) in the hope that he would break up the I1 block but he didn't show up as confirmed negative for any of them - he was positive for ~270 of them with the remainder being no calls.

But all possible - in the last couple of years within I-Z140 we had a couple of guys test Z140+/Z141- and then another broke up several of the phyloequivalent SNPs on the I-F2642/S2169 branch which caused a bit of a reordering of the tree. But I thought if someone would break up the I1 block, would have found them by now. Wait and see, I guess.

You can see how many in your own Big Y test are covered if you go to your YFull homepage and click on YReport on the left hand menu - it will open a new window and the SNPs will be colour-coded on the YReport - green for confirmed positive, red for negative, yellow for ambiguous, grey for no call (the grey is a little harder to read since it is not too diferent from the green colour). I don't have a Big Y myself to look at but I have a FGC YElite - this has three ambiguous calls and 38 no calls in the big I1 block. That's not too bad and comparable to the I-CTS12798 individual mentioned above. I also have a 30x WGS which gets almost everything - zero no calls in the big I1 block and two ambiguous calls. I'd expect that a Big Y Y500 would have less coverage than both of these and a Big Y Y700 to be similar to the YElite.

I know for sure there are several WGS of I1 folks down the I-DF29 branch. I assume that the ambiguous and no calls in the Y-enrichment tests like Big Y are inferred positive. I guess they could be verified for sure with a Sanger test at YSEQ. I have no idea about the I-Z17954 branch and I think the two folks at I-CTS12798 are a Big Y and a YElite. But most people probably don't look at that part of their results as they are more concerned with their terminal branch and novel variants.

I agree that the using the I1 label for ancient samples like this is rather misleading - unfortunately they have been quoted as such many, many times and most people reference the blog posts and secondary analysis rather than the primary data. For several of those samples, the lack of data makes it impossible to assign with certainty. I try calling them pre-I1 but I'm not sure if that's appropriate either as it implies that they are an ancestor of modern I1 while to me it's not clear if they are an intermediate or a dead lineage. Not enough data to confirm or disconfirm.

Thanks for the detail response. I checked out my YReport for my Big Y-500 test and upon an initial quick scanning I am positive for over 200 SNPs associated with I1 and am ambiguous for 4 SNPs. I agree with you now that enough SNPs are tested with the nextgen tests to confidently classify people as I1.

[deadly77]

For the ancient samples where I'm reliably able to assign a subclade (even if it's only as far as I-DF29), I haven't dug too heavily into the big block of I1 SNPs as in those cases they're all inferred to be positive. The only bit of data that I have came from when I was having a chat with one of the admins of the I-L22 project a while ago - he was kind enough to share the YFull YReport for the highest quality Icelandic sample HSJ-A1. HSJ-A1 had 13 SNPs that were no call from the big I1 block. Although that's helped by the testing done on HSJ-A1 that was a WGS.

[oz]

Deadly77 what do you think about this BAL051 finding pushing back the TMRCA of I1 would that make sense? Does it have enough positive reads or however that works. Currently it's estimated at around 5,000 ybp isn't it? While DF29 and his descendants should obviously remain estimated to be much younger.

[deadly77]

I don't think it would push back the TMRCA of I1 per the YFull tree estimate since that's based on counting the SNPs of modern testers backwards to the common ancestor branch, albeit with a few qualifiers and correction factors - see https://www.yfull.com/faq/what-yfull...n-methodology/ and https://www.yfull.com/faq/how-does-y...-tmrca-and-ci/ for details and you can see the numbers that make up the calculation if you click on the white info button next to the I1 branch https://www.yfull.com/tree/I1/ - there's a range of age estimates based on samples on that branch and ages of downstream branches and they all get averaged to a current TMRCA 4600 years ago, although there is a 95% confidence interval for 4000-5100 years ago. I don't think the TMRCA of I1 gets pushed back unless a modern tester (or prefereably multiple testers) is shown to be definitively negative for some of the SNPs in the Big I1 block.

What I think that you're actually asking is that if you insert BAL051 into the I1 block, then there's a new branch point in between I and I1 - that would have it's own formed date and TMRCA for that branch. I think it's hard to calculate for BAL051 (and same for SF11) because there is a lot of no calls for BAL051 in the big I1 block. Someone on the I-Z140 facebook page has postulated a age estimation for the branch point at BAL051 at about 18500 years ago by applying the ratio of 20/33 (ancestral/total read SNPs) to the years between the formed date and TMRCA. I'm not sure if I agree - there's no guarantee that same ratio of ancestral and derived SNPs is consistent among the 279 SNPs that were not read, so that's a big assumption with a significant amount of error. Also, not all of the 312 SNPs on the I1 big block are used for age estimation (which is why I included references to the YFull FAQ regarding SNPs in the CombBED region, etc.) - on the YFull tree it looks like it ranges between 142 and 192 SNPs depending on sample, so there would need to be a correction factor for that as well. I think it's a bit more complicated so I'll think about it a bit more, but I think it's going to come down to not enough data to get a confident answer based on the no calls.

[oz]

Right I figured your response would be this line of thinking. I just figured I'd bring up the question since I don't think anyone else has so far. And the reason I asked if it's possible to consider this as proof of the earliest I1-M253 because on some other threads and including this academic study they're labeling it as I1 when technically that's not what it is if there's not enough positive snps, not even close really and therein lies the confusion. It should be labeled as pre I1 or I* but they're still calling it I1.

Either way I wish it was positive for all the SNPs and resembled more the modern I1s. I kinda always felt weird that I belong to a haplogroup that seems to have appeared out of nowhere much more recently than all the other major HGs, that's just me. The fact that I1 has such a large number of defining SNPs is also pretty strange in itself. But I'm not any kind of expert on this science to be able to even hypothesize what that means.

[spruithean]

Right I figured your response would be this line of thinking. I just figured I'd bring up the question since I don't think anyone else has so far. And the reason I asked if it's possible to consider this as proof of the earliest I1-M253 because on some other threads and including this academic study they're labeling it as I1 when technically that's not what it is if there's not enough positive snps, not even close really and therein lies the confusion. It should be labeled as pre I1 or I* but they're still calling it I1.

Either way I wish it was positive for all the SNPs and resembled more the modern I1s. I kinda always felt weird that I belong to a haplogroup that seems to have appeared out of nowhere much more recently than all the other major HGs, that's just me. The fact that I1 has such a large number of defining SNPs is also pretty strange in itself. But I'm not any kind of expert on this science to be able to even hypothesize what that means.

I share this sentiment. With the amount of people who belong to the I-M253 umbrella, there sure isn't a whole lot of ancient samples that are far removed from the eras of the current known ancient I1 individuals. That bottleneck must have been intense.

[oz]

I share this sentiment. With the amount of people who belong to the I-M253 umbrella, there sure isn't a whole lot of ancient samples that are far removed from the eras of the current known ancient I1 individuals. That bottleneck must have been intense.

Well when for tens of thousands of years you're constantly inbreeding and hiding like Bigfoot I guess that's what happens.

That may sound like a joke but it's what the academics basically claim.

[deadly77]

Right I figured your response would be this line of thinking. I just figured I'd bring up the question since I don't think anyone else has so far. And the reason I asked if it's possible to consider this as proof of the earliest I1-M253 because on some other threads and including this academic study they're labeling it as I1 when technically that's not what it is if there's not enough positive snps, not even close really and therein lies the confusion. It should be labeled as pre I1 or I* but they're still calling it I1.

Either way I wish it was positive for all the SNPs and resembled more the modern I1s. I kinda always felt weird that I belong to a haplogroup that seems to have appeared out of nowhere much more recently than all the other major HGs, that's just me. The fact that I1 has such a large number of defining SNPs is also pretty strange in itself. But I'm not any kind of expert on this science to be able to even hypothesize what that means.

In my opinion, I would not say that BAL051 should be considered the earliest example of I1. For me, the I1 haplogroup is defined by that big block of >300 SNPs and any sample that is ancestral for some of the SNPs in the big block isn't fully I1 but a proto lineage whose descendants may have evolved into fully I1, or they may be representatives of extinct lineages where the descendants died out and not found today in the modern population. Richard Miller made a chart which I think describes this nicely:
I1 age estimation.jpg
I'd include SF11 in that category of not fully I1 due to the ancestral I1 SNPs found in that sample. Not unexpected as these two samples are dated prior to the modern I1 TMRCA. I'd also include the four RISE samples from Allentoft and BAB5 in this category of not fully I1 due to a lack of reinforcing positive evidence since those designations are made on a very small number of SNPs. For me, the data is too shaky to confirm these samples as fully I1. This is just my opinion, and I know for sure that it isn't shared by everyone. I agree that it does add to the confusion. It is interesting that the authors of the paper (I'll call it Haak since that's the corresponding author) refer to BAL051 as I1, while the other big recent paper on the Iberian peninsula by Olade et al has a fair few samples that are listed as I, with some confirmed as I(xI2a1,xI2a2,xI2c) in Table S4 but they don't designate any as I1. It'll be interesting to have a closer look at some of the samples in the Olade paper once the BAM files are uploaded to ENA. Looks like the BAM files from that study will be deposited here (nothing there yet): https://www.ebi.ac.uk/ena/data/view/PRJEB30874

Yeah, the large number of defining SNPs is down to an unbroken line of descent from about 27,500 years ago to about 4,600 years ago when we start to see expansion and diversity of subclades. As Spruithean says, the bottleneck must have been intense. It did remind me of a snippet I heard when I was listening to The Insight podcast (well worth listening to in my opinion) last year - went back and found it on their website here https://insitome.libsyn.com/y-chromosomal-stars - there's a few comments from Razib Khan about I1 starting about 36 and a half minutes into the podcasts where he theorizes that as the Steppe people moved into Europe, they assimilated some of the local indigenous Europeans, with I1 "hitching a ride" among a tribe that was mostly non-I1.

I think it's impressive that an I1 lineage survived so long in the midst of that competition without dying or daughtering out, long enough to survive to get to the end of the bottleneck and propagate and expand from there.

[oz]

Indeed it is impressive when you look at it from the survival aspect. It was a very headstrong, resilient I* that managed to keep a small lineage going for so long before branching out not too long ago amongst so many other lines who weren't as lucky.

[deadly77]

Trying to scrape the barrel on the low coverage pre-I1 samples and look for SNPs that are read in more than one sample. It's not an exhaustive list:

CTS3506/Z2765: BAL051 ancestral (1A), RISE179 derived (1G).
FGC2433: BAL051 ancestral (1G), SF11 derived (1A).
Z2806/CTS6765: BAL051 ancestral (1T), RISE210 derived (1C).

So at least no red flags that wouldn't be easily explainable such as a derived allele for the older BAL051 sample and a corresponding ancestral allele on the same SNP in one of the younger SF11 or RISE samples. But a really small dataset. Not really sure if this means anything.