VCF to BCF conversion error: [E::bcf_hdr_add_sample] Duplicated sample name

[LeonDK]

In converting a .vcf file to a .bcf file a get the following error message:

[E::bcf_hdr_add_sample] Duplicated sample name

I am running conversion using bcftools like so:

bcftools convert -O u -o /path/to/input/file.bcf  /path/to/output/file.vcf.gz

How do I fix this?

[jmarshall]

Sample names are parsed from the #CHROM POS ID REF line, and can't appear more than once (at least for bcftools's implementation; not sure if this is clear in the VCF spec).

The actual "Duplicated sample name" error message also tells you which sample name it's complaining about.

[LeonDK]

Sample names are parsed from the #CHROM POS ID REF line, and can't appear more than once (at least for bcftools's implementation; not sure if this is clear in the VCF spec).

The actual "Duplicated sample name" error message also tells you which sample name it's complaining about.

Yes, I was looking for a tool, which could identify and remove sample duplicates

[jmarshall]

That error message identifies duplicate sample names, or to find them all at once:

bcftools view -h file.bcf|grep '^#CHROM'| cut -f10-|tr '\t' '\n'|sort|uniq -c|grep -v '^ *1 '

Then you could e.g. hack the #CHROM line in a text editor and use bcftools view -s to remove the duplicates you've marked via the text editor. Or use awk or similar.

[LeonDK]

That error message identifies duplicate sample names, or to find them all at once:

bcftools view -h file.bcf|grep '^#CHROM'| cut -f10-|tr '\t' '\n'|sort|uniq -c|grep -v '^ *1 '

Then you could e.g. hack the #CHROM line in a text editor and use bcftools view -s to remove the duplicates you've marked via the text editor. Or use awk or similar.

I wrote me a small script to search and identify sample name duplicates and exclude, such that only one is kept for each possible duplicate.

Thanks for input.