atks / vt

A tool set for short variant discovery in genetic sequence data.
http://genome.sph.umich.edu/wiki/vt
MIT License
190 stars 3 forks source link

smart decomposition does not export AD for multiallelic variant #88

Closed wangshu84 closed 6 years ago

wangshu84 commented 6 years ago

Hello, I used smart decomposition for getting AD and GQ. However, decomposed output have "." for all multiallelic variants. It is VCFv4.2 with FORMAT=<ID=AD,Number=R...

e.g. For a multiallelic variant OLD_MULTIALLELIC=chr1:115256325:A/G/T GT:AD:DP:GQ:PGT:PID:PL 0/0:.:15:45:.:.:0,45,632,45

if not multiallelic: GT:AD:DP:GQ:PL 0/0:15,0,0:15:45:0,45,648,45,648,648

1).Wonder is there a way to get all AD information for all alt alleles. 2). For multiallelic situation, e.g. 2 alt alleles, is there a way to write entry for new genotype GT=1/2

Thank you!

atks commented 6 years ago

@wangshu84

Can you give me an example VCF to work with?

wangshu84 commented 6 years ago

@atks Attached is a test.vcf file. It has extension .txt just for uploading. Thank you! vt_test.vcf.txt

atks commented 6 years ago

@wangshu84

apologies for the late reply, can you try it without the -s option and see if that is what you want?

also, may I know where was the VCF derived from?

atks commented 6 years ago

@wangshu84

2). For multiallelic situation, e.g. 2 alt alleles, is there a way to write entry for new genotype GT=1/2 can you explain abit more about this.

apologies for the late replies.

wangshu84 commented 6 years ago

@atks unfortunately without -s option, it does not output AD neither.

atks commented 6 years ago

@wangshu84 - apologies for the delay, I can only respond on this around next week.

wangshu84 commented 6 years ago

@atks looks like header INFO field is the reason causing missing AD for multiallelic variants. If change format of AD field to "Number=." (input vcf has "Number=R"), the smart decompose will output AD values. The input vcf is by GATK_3.7

atks commented 6 years ago

@wangshu84 apologies about this, I should have spotted it much earlier.