konradjk / exac_browser

Browser for ExAC consortium data
http://exac.broadinstitute.org
MIT License
106 stars 54 forks source link

chr X warning #327

Open rjsicko opened 6 years ago

rjsicko commented 6 years ago

I think the warning "This means that the site is covered in fewer than 80% of the individuals in ExAC, which may indicate a low-quality site." is not account for male/female status in AC.

example variant with the warning

rational for why I think that filter isn't accounting for male/female:

Number of individuals in ExAC 33,644 - male 27,062 - female 60,706 – total

So, there should be 87,768 alleles for an X chr SNP not in the PAR.

For the 80% individuals cutoff we can have the two extremes: 48,564 individuals – 33,644 males = 14,920 females -> 63,484 alleles as a cutoff for 80% individuals (if all the 20% fails are female)

48,564 individuals – 27,062 females = 21,502 males -> 75,626 alleles as a cutoff for 80% individuals (if all the 20% fails are male)

edit: related issue reported in gnomad repo.

line 175 of 4b9e607 exac_browser/templates/variant.html line 428 of 07d003b gnomad_browser/templates/variant.html (commented out) {% if variant.allele_num < 97129.6 %}

rjsicko commented 6 years ago

PAR coordinates for GRCh37.p13 in case it helps when adjusting filter.

Name | Chr | Start | Stop PAR#1 | X | 60,001 | 2,699,520 PAR#2 | X | 154,931,044 | 155,260,560 PAR#1 | Y | 10,001 | 2,649,520 PAR#2 | Y | 59,034,050 | 59,363,566