PHAX5574 discussion

[Norfern-Ostrobothnian]

Yes, women have twice the X-chromosomes thus possess twice the chance of passing on a haplotype to the next generation. One thing to note is that so far all new variants have been based on male samples only with the exception of Neanderthals and Denisovans, in which only homozygous locations were used. Due to the nature of the data it is very hard to resolve where a new allele might belong if the female sample has evidence of possessing two distinct haplotypes. Although due to the prevalence of haplotypes such as B, it might be possible to ascertain new haplotypes from females as well if the region is otherwise homozygous, or in the case where the new allele is also homozygous. For robustness sake I will continue to use males until there's enough so that females could easily be assigned both of their haplotypes.

[Fitis]

Yes, women have twice the X-chromosomes thus possess twice the chance of passing on a haplotype to the next generation. One thing to note is that so far all new variants have been based on male samples only with the exception of Neanderthals and Denisovans, in which only homozygous locations were used. Due to the nature of the data it is very hard to resolve where a new allele might belong if the female sample has evidence of possessing two distinct haplotypes. Although due to the prevalence of haplotypes such as B, it might be possible to ascertain new haplotypes from females as well if the region is otherwise homozygous, or in the case where the new allele is also homozygous. For robustness sake I will continue to use males until there's enough so that females could easily be assigned both of their haplotypes.

With several family members, it is also possible to reconstruct the haplotype of heterozygous women (phasing) or even without family data using long-read sequencing (but of course still not very accessible). I found several SNPs not listed in your Excel table in my family data on PHAX5574 and quite rare in populations (most <5% in the GnomAD v4.1 population database but not private to my family), perhaps these SNPs characterise sub-branches of B-h1 and C-h2.

[Norfern-Ostrobothnian]

Almost all subclades in this haplotype are defined by a single mutation. That is also why using ancient DNA is though, due to the inherent noise in the data.

[Kale]

On Ancestry and 23andme V5 the only snp covered (among those listed in PHAX5574 haplotypes.xlsx ) is...
B
rs3131400

Luckily I have that derived, so at least I know I'm B

[ronin92]

C
rs12557363
B
rs3131400
H
rs148852196

rs3131400 is derived in B and E.  E appears to be determined by X:94554411 (which is rs141198488, not annotated as such in xlsx) at which it is derived while B is ancestral.  ~1% allele frequency.

I appear to be a B with a further SNP reported in dbSNP 155 but the derived allele was not observed in the 14050 samples they have.  They don't say the basis on which they added it - must have observed it somewhere :-(

[ChrisR]

I did manual analysis on a male WGS result which seems to have 10 variants in PHAX5574.
One 94561636 A>G is not in line with the h63 F finding, another 94583459 C>G seems not covered yet or novel.

Coordinates hg19: chrX:94544032-94582132
Coordinates hg38 UCSC LiftOver: chrX:95289033-95327133

WGS ~60x hg38 VCF
chrX:95291634 G>A
chrX:95292855 A>C
chrX:95295134 C>G
chrX:95298214 A>T
chrX:95298727 T>A
chrX:95301587 G>A
chrX:95306637 A>G
chrX:95312922 G>A
chrX:95326715 G>A
chrX:95328460 C>G

hg19 UCSC LiftOver
chrX:94546633 G>A rs866421262 h63 F in PHAX5574 haplotypes.xlsx
chrX:94547854 A>C rs3122217 h63 F in PHAX5574 haplotypes.xlsx
chrX:94550133 C>G rs12557363 h63 F in PHAX5574 haplotypes.xlsx
chrX:94553213 A>T rs3122216 h63 F in PHAX5574 haplotypes.xlsx
chrX:94553726 T>A - h63 F in PHAX5574 haplotypes.xlsx
chrX:94556586 G>A rs3131397 h63 F in PHAX5574 haplotypes.xlsx
chrX:94561636 A>G - (not h63 F in PHAX5574 haplotypes.xlsx)
chrX:94567921 G>A rs5949556 h63 F in PHAX5574 haplotypes.xlsx
chrX:94581714 G>A - h63 F in PHAX5574 haplotypes.xlsx
chrX:94583459 C>G (not in PHAX5574 haplotypes.xlsx)

[Norfern-Ostrobothnian]

In the next edition of the haplotype spreadsheet I should finish the SNP annotations and add GRCh38 coordinates as well.

[TanTin]

I have a special interest in CHR X (23).
However we don't have these snips in D . Reich's dataset.

I found only one:
23
rs111514686
1.19655
94558358
T
A

So for these who want to confirm their group they should check in the raw data to find some matching snips. That means lot of work to do to find out and clasify the existing ancient samples.

[ChrisR]

However we don't have these snips in D . Reich's dataset.

Thx for notifying. I wonder why that is, since the inheritance pattern of X allows some more female ancestral side conclusions.
This also begs for some kind of automated solution in analysis and haplotype tree building. Maybe YFull is interested?

[TanTin]

Thx for notifying. I wonder why that is, since the inheritance pattern of X allows some more female ancestral side conclusions.
This also begs for some kind of automated solution in analysis and haplotype tree building. Maybe YFull is interested?

X-chromosome has 154,259,566 bp .
From these 154 millions only 50k are included in 1240k dataset. The selection of these 1240k was done many years ago.
The Y -chromosome is near 3x times smaller than the X . (The Y chromosome spans more than 59 million base pairs) . Only 33 k of the Y-CHR are included in the 1240k selection.