This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Combining VCF files using GATK

I am trying to combine vcf file generated after haplotypecaller from different variety of breed. The total number of samples is ~100. I ran the command

gatk CombineGVCFs -R genome.fna --variant 1.vcf.gz --variant 2.vcf.gz.....variant97.vcf.gz -O combine.g.vcf.gz

but when I am trying to check the file it contain only 41 samples instead of 97 in vcf. If anyone know what I am doing wrong or any experience please share how to fix such situation.

wgs vcf gwas
• 5,073 views • link
updated by niyomiw88 • niyomiw88 • written by tothepoint tothepoint

Which version of GATK are you using?

• 0 views • link
written by Arup Ghosh Arup Ghosh

I am using gatk4.1.4.0

• 0 views • link
written by tothepoint tothepoint

Have you checked the log file to make sure the command completed. Also I'll recommend GenomicsDBImport instead of CombineGVCFs.

• 0 views • link
written by Arup Ghosh Arup Ghosh

I checked the log file and there was no such issue. I already gave one more shot to CombineGVCFs but more curious to check using GenomicsDBImport. Thanks

• 0 views • link
written by tothepoint tothepoint

I would check which samples are missing, does this give you a clue? Perhaps the index is missing from the missing vcf files. Check which samples have been included:

zgrep -m 1 '^#CHROM' combine.g.vcf.gz | head -n 1 | cut -f 10- | tr '\t' '\n'
• 0 views • link
written by rbagnall rbagnall

Were you able to figure out what went wrong? I have a similar issue with CombineGVCFs. All my samples are literally merged into ONE. I have 243 individuals, but the final VCF shows all the variants as in one individual. I have no idea what is going on. I checked the log file. It looks fine. Reading in all the individual samples with no errors. Here is the command i used:

gatk --java-options "-Xmx30G" CombineGVCFs -R reference --variant file1.g.vcf  --variant file2.g.vcf ... -O cohort.g.vcf

Any tips on how to fix this? I can't use the final vcf when there are no information on individuals. Thanks in advance

• 0 views • link
written by niyomiw88 • niyomiw88

Only add an answer if you're answering the top level question. If you have a follow up or "I have this problem too" statement, use Comments instead. I've moved your post to a comment this time.

• 0 views • link
written by Ram Ram

There was indexing issue from the missing combining files. I cross checked and indexed those file with 

bgzip -c missing_vcf.vcf > missing_vcf.vcf.gz
tabix -p vcf missing_vcf.vcf.gz

And re-run them with all file combined successfully.

• 0 views • link
written by tothepoint tothepoint

0 answers

No answers yet.

Log in to answer this question.

More posts like this