search

zhanxw / rvtests

Rare variant test software for next generation sequencing data
129 stars 41 forks source link

Issue with processing Chromosome X summary files #99

Open jdarias93 opened 4 years ago

jdarias93 commented 4 years ago

Dear RvTests team,

We are working with the chromosome X result files that were generated with Rvtests and submitted to the GIANT consortium for meta-analysis. We noticed that many of chr X results files have values of ‘0’ for the columns INFORMATIVE_ALT_AC, CALL_RATE, HWE_PVALUE, N_REF, N_HET, and N_ALT for all SNPs (even though all the SNPs were imputed). I’ve attached the first 1000 lines from an output file we’re working with.

We’ve noticed is that this is more likely to happen in Chromosome X files with male subjects only, but we have also seen it with a few files including women. Another pattern is that this usually affects the entire file (or all but one or a few SNPs). Do you know why this is occurring with RVtest for chromosome X and is there a way to correct it? Also, do you think the summary stats are still valid?

Please let us know if you need any more information. Thank you in advance, Joshua Arias

SampleRvTests_first1000lines.txt

medinacarolina commented 2 years ago

Hi Joshua was wondering if you solved your issue

jdarias93 commented 2 years ago

Hello, I have not, but we are trying a few workarounds, namely: 1) Setting all HWE p to 1 in order to bypass filtering during a downstream analysis (RAREMETAL) 2) Replacing all "NA" AF values with those from the study's respective .INFO file 3) Manually removing all variants with a p-value of 0 or 1

-Josh

On Wed, Oct 6, 2021 at 5:23 AM medinacarolina @.***> wrote:

Hi Joshua was wondering if you solved your issue

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/zhanxw/rvtests/issues/99#issuecomment-935826098, or unsubscribe https://github.com/notifications/unsubscribe-auth/AG6HSXTOWPYGHY42LAYYJSDUFQIR3ANCNFSM4M23JUPA . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

Haojie-Lu commented 2 years ago

Hi Joshua and Carolina.

I also faced this problem recently and found ways to solve it.

From my side, I found the problem is caused by the gender code. (In my input .ped files, male is 0, female is 1. while in RVtests, male should be 1 and female should be 2). In my log files, some samples are recognized as male, some are sex unknown, and the problem happened. When I recode gender as 1 and 2 in my ped file, there is no more SNP has the AF of NA.

Hope this can help you.

Best

Haojie