Missing samples in the output vcf file created using GenotypeGVCFs in GATK

Question

[kk.mahsa]

Hi everyone

I used the following method to create a VCF file with 50 samples.

For each sample

 java -jar gatk_3.7-0/GenomeAnalysisTK.jar -T HaplotypeCaller -R Ref.fasta -I input.bam -o output.g.vcf.gz -ERC GVCF

and then for all samples

java -Xmx64G -jar gatk_3.7-0/GenomeAnalysisTK.jar -T GenotypeGVCFs -R Ref.fasta -V output1.g.vcf.gz -V output2.g.vcf.gz ... output50.g.vcf.gz -o 50_samples.vcf.gz

But my final VCF file has 45 samples instead of 50 samples. I don't get any error messages.

Can anyone help me to solve this problem?

Answer 1

[Pierre Lindenbaum]

check you have 50 disctinct samples: what is the output of:

for F in output*.g.vcf ; do bcfools query -l "${F}" ; done | sort | uniq | cat -n 

[kk.mahsa]

I ran your command on the 15 samples (Sam1 to Sam15) that the missing data belonged to them and the command output was:

1 sam1

2 sam10

3 sam11

4 sam12

5 sam13

6 sam14

7 sam15

8 sam2

9 sam3

10 sam7

11 sam8

12 sam9

what is that mean?

[Pierre Lindenbaum]

what is that mean

you only have 19 distinct samples in all your VCF. . My guess is that the original BAMs have the same samples. Check with

cat list.of.path.to.bam | samtools samples

[kk.mahsa]

Thanks dear Pierre for your help; I checked the BAM files and realized that I made a mistake in adding the sample name to the BAM files and it is a duplicate name