brentp
commented
3 years ago It must be that the AD or DP field is not present in the samples fields for some variants
Closed sprakashUTH closed 3 years ago
brentp
commented
3 years ago It must be that the AD or DP field is not present in the samples fields for some variants
sprakashUTH
commented
3 years ago It looks like old multiallelic variants that were split caused the problem. Is there any workaround? Also, the tool continues to output only zeroes (no variants) in all categories except for X chromosome variants. I'm not getting any other errors. This time I used a jointly called vcf that works for other analyses. Do you have any suggestions?
brentp
commented
3 years ago was it called with GATK? can you share the header of the VCF?
sprakashUTH
commented
3 years ago brentp
commented
3 years ago I don't see any problem with the header.
But there doesn't appear to be any CSQ information. If you are using INFO.impactful that is required. You can add that with snpEff, VEP, or bcftools csq
brentp
commented
3 years ago also, what did you use to decompose and normalize?
sprakashUTH
commented
3 years ago sprakashUTH
commented
3 years ago vt
brentp
commented
3 years ago I guess you could try using bcftool norm instead. The other problem is that slivar should still call some variants. What does your pedigree file look like? You have affected samples?
sprakashUTH
commented
3 years ago Yes I have 42 trios and more than 100 affected individuals
brentp
commented
3 years ago if the parents are affected, then you'd use autosomal dominant. the rare disease page lists only recessive and denovo.
sprakashUTH
commented
3 years ago No, they are trios. Parents are unaffected. Slivar isn’t working.
From: Brent Pedersen @.> Date: Friday, April 16, 2021 at 6:37 AM To: brentp/slivar @.> Cc: Prakash, Siddharth K @.>, Author @.> Subject: Re: [brentp/slivar] Errors with rare variant expression (#87)
EXTERNAL EMAIL
if the parents are affected, then you'd use autosomal dominant. the rare disease page lists only recessive and denovo.
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_brentp_slivar_issues_87-23issuecomment-2D821114504&d=DwMCaQ&c=bKRySV-ouEg_AT-w2QWsTdd9X__KYh9Eq2fdmQDVZgw&r=_wR25Q6_6V5aHtne4gUQAZTHcU0BjRWiyj5K1TTKqYU&m=8Rc9Vgz_DJfnNI2HuTW4kceQQp2XJSLPbI_OvzSgX5o&s=16vETCuR4JXqeFMJCQcm2xx6EsJJfrzAwNMjqW33JqU&e=, or unsubscribehttps://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_notifications_unsubscribe-2Dauth_API7HLX2TLXDEEGHOWOY7CTTJAORHANCNFSM43AIJVYA&d=DwMCaQ&c=bKRySV-ouEg_AT-w2QWsTdd9X__KYh9Eq2fdmQDVZgw&r=_wR25Q6_6V5aHtne4gUQAZTHcU0BjRWiyj5K1TTKqYU&m=8Rc9Vgz_DJfnNI2HuTW4kceQQp2XJSLPbI_OvzSgX5o&s=BXPlVKssRLslgJszCaIgKJtrba1wVgA0wzge97IGP70&e=.
sprakashUTH
commented
3 years ago No, they are trios. Parents are unaffected. Slivar isn’t working.
brentp
commented
3 years ago If you can provide a pedigree file and a vcf of relevant samples with a single variant where slivar "isn't working" then I can look into it. I suspect something is wrong with your data but it's hard to guess with the little and changing information you are giving. If you can't provide that, then please close the issue.
As to the warnings about missing DP. I think that might be some problem with vt but again, I don't have enough information to help you more on that.
sprakashUTH
commented
3 years ago The information never changed. I input a vcf file with 500,000 variants and slivar outputs 0 variants and gives the errors I showed you. I can send you the entire vcf and the pedigree if you want and you can test it for yourself.
brentp
commented
3 years ago that would be great.
sprakashUTH
commented
3 years ago OK, please send a private link to upload.
brentp
commented
3 years ago can you email me? bpederse <| at |> gmail
brentp
commented
3 years ago I can recreate the problem. I hope to have a fix out this week. Thanks very much for providing the data to recreate the issue!
brentp
commented
3 years ago So, this is really a problem with your VCF. There is not a single het variant (or at least very few) with DP > 0. The AD field will have, e.g. 12,14, but for that sample and variant, the DP field will be .
I have updated slivar-functions.js to avoid using DP and rely only on AD. So if you update your slivar-functions.js, then you'll get some results.
Hope this helps. Thanks again for providing the test-case.
sprakashUTH
commented
3 years ago Thanks! I tried it and it works. I got the information I need out of slivar.
Command: ./slivar expr --vcf $vcf --ped $ped --pass-only -g $gnomad --info 'INFO.impactful && INFO.gnomad_popmax_af < 0.01 && variant.FILTER == "PASS" && variant.ALT[0] != "*"' --js slivar-functions.js --family-expr 'denovo:fam.every(segregating_denovo) && INFO.gnomad_popmax_af < 0.001' --family-expr 'recessive:fam.every(segregating_recessive)' --family-expr 'x_denovo:(variant.CHROM == "X" || variant.CHROM == "chrX") && fam.every(segregating_denovo_x) && INFO.gnomad_popmax_af < 0.001' --family-expr 'x_recessive:(variant.CHROM == "X" || variant.CHROM == "chrX") && fam.every(segregating_recessive_x)' --trio 'comphet_side:comphet_side(kid, mom, dad) && INFO.gnomad_nhomalt < 10' | bcftools csq -s - --ncsq 40 -g $gff -l -f $fasta - -o $cohort.vcf
Error (recurred hundreds of times): [slivar] javascript error. this can some times happen when a field is missing. error from duktape: unknown attribute:DP for expression:fam.every(segregating_denovo) && INFO.gnomad_popmax_af < 0.001
I decomposed and normalized the vcf. What causes these errors? Should these errors affect my confidence in the output?