Problem of bcftools merge TCGA vcf files

hi, everyone I am college students and try to do case-control GWAS through TCGA vcf data. i use gdc-client download TCGA vcfs, and i got many single-samples vcf . Therefore, i attempt to merge single-samples vcf into multiples by bcftools. I encounter problems below and i search for some methods on net but no idea.Please help or try to give some ideas how to solve this. this is my commands:

bcftools merge  *.vcf.gz -Oz -o merge.vcf.gz

The Error is : Duplicate sample names (NORMAL), use --force-samples to proceed anyway. so i

bcftools merge --force-samples *.vcf.gz -Oz -o merge.vcf.gz

but i lost most of chrom info in merge.vcf . and try to see the header

zgrep -n '^#CHROM'  merge.vcf.gz |cat

the outcomes are like

CHROM   POS ID  REF ALT QUAL    FILTER  INFO    FORMAT  NORMAL  TUMOR   2:NORMAL    2:TUMOR 3:NORMAL    3:TUMOR 4:NORMAL    4:TUMOR 5:NORMAL    5:TUMOR

In the end ,thank in advance and sorry for my poor english.

snp merge

• 2,707 views • link

updated 5.9 years ago by Jorge Amigo

• written 5.9 years ago by l66081129 •

1 answer

If your vcf samples are all internally named "NORMAL" and "TUMOR", my first suggestion would be to rename them before merging them:

for file in $(ls *.vcf.gz | grep -v renamed | grep -v merge); do
 bcftools reheader -s <(echo ${file/.vcf.gz}) -o ${file/.vcf.gz/.renamed.vcf.gz} $file
 tabix -fp vcf ${file/.vcf.gz/.renamed.vcf.gz}
done

You wouldn't have any problem to merge them afterwards:

bcftools merge -Oz -o merge.vcf.gz *.renamed.vcf.gz

You should see all the correct sample names on the last header line

zgrep ^#CHROM merge.vcf.gz

• 0 views • link

written 5.9 years ago by Jorge Amigo

thank you for your answer sincerely !! it helps me a lot i,ve tried just now but there is error ,so i take step by step

for file in $(ls *.vcf.gz |grep -v rename |grep -v merge);do bcftools reheader -s ${file/.vcf.gz} -Oz -o ${file/.vcf.gz/.renamed.vcf.gz} $file ;done  2>>error

and error seems like bcftools reheader -s ${file/.vcf.gz} doesn't load vcf header in . i would try to learn and adjust it it suggest me use bcftools view -h old.bcf > header.txt and bcftools reheader -h header.txt

thanks again , have a nice day

• 0 views • link

written 5.9 years ago by l66081129 •

Sorry I didn't test the reheader command before. It doesn't accept a sample name but a file containing sample names, nor it doesn't accept the -Oz common bcftools option, so I've edited the code and it now works flawlessly.

• 0 views • link

written 5.9 years ago by Jorge Amigo