Male heterozygotes on X chromosome
Hi all,
I am working with the WES data set in Hail, on a gene located on the X chromosome.
Before quality control, I have 2.5k sites all in a non-pseudoautosomal region (checked with mt_x_m.locus.in_x_nonpar()).
Calls from deepvariant are diploid:
print(mt_x_m_par.aggregate_entries(hl.agg.counter(mt_x_m_par.GT.ploidy)))
{2: 335236742, None: 4499900}
But I still get quite a few heterozygotes:
print(mt_x_m.aggregate_entries(hl.agg.counter(mt_x_m.GT.is_het())))
{False: 335138969, True: 97773, None: 4499900}
print(mt_x_m.aggregate_entries(hl.agg.counter(mt_x_m.GT.is_hom_var())))
{False: 335124731, True: 112011, None: 4499900}
What might explain these het calls (which should not happen)? Would it be best to remove these GT?
Thanks in advance,
Comments
2 comments
One of our Bioinformaticians has provided the following comments:
We have previously checked that there hasn’t been a mislabelling issue. One hypothesis is that the variant calling procedure may have produced such an artefact in the data – could you check whether any of these heterozygous variant calls are actually ‘-/0’ or ‘-/1’? Alternatively, does the number of errant samples correspond to the proportion of the population for whom recorded sex differs from genetic sex? This is explored in more detail in the following publication: https://www.nature.com/articles/s41586-018-0579-z. However, we cannot answer this point definitively and would otherwise direct you to the authors of the work which produced the data (see: https://biobank.ctsu.ox.ac.uk/showcase/label.cgi?id=170).
Hi,
I also noticed some heterozygous males in non-PAR chrX using the WES data. Not sure if this is the case, but it might have to do with a bug in SHAPEIT5 when phasing, where missing calls are incorrectly assigned a heterozygous genotype (https://github.com/odelaneau/shapeit5/issues/100).
Please sign in to leave a comment.