htsjdk.tribble.TribbleException: The provided VCF file is malformed at approximately line number 92283: Duplicate allele added to VariantContext: C, for input source:

Question

[DareDevil]

I run variantsToTable module of gatk as follows:

gatk VariantsToTable -V 001.vcf -F CHROM -F POS -F ID -F REF -F ALT -F QUAL -F FILTER -F INFO -F FORMAT -F Father -F Mother -F Child -F ADP -F STATUS -F CSQ  -GF GT -GF GQ -GF SDP -GF DP -GF RD -GF AD -GF FREQ -GF PVAL -GF RBQ -GF ABQ -GF RDF -GF RDR -GF ADF -GF ADR -O 001.table

It has shown an error as follows:

htsjdk.tribble.TribbleException: The provided VCF file is malformed at approximately line number 92283: Duplicate allele added to VariantContext: C, for input source: 001.vcf

I checked what is there in the line number

sed -n 92283p 001.vcf

the output was:

chr1    35816798    .   CAAAAAAAAAAAAA  C,C .   PASS    ADP=20;STATUS=2;CSQ=-|intron_variant|MODIFIER|AGO4|ENSG00000134698|Transcript|ENST00000373210|protein_coding||1/17||||||||||1||HGNC|HGNC:18424|YES|CCDS397.1||  GT:GQ:SDP:DP:RD:AD:FREQ:PVAL:RBQ:ABQ:RDF:RDR:ADF:ADR    0/1:51:22:22:7:13:59.09%:6.4422E-6:38:47:6:1:13:0   0/2:20:27:27:13:6:24%:9.828E-3:38:39:13:0:6:0   0/2:22:13:13:5:6:46.15%:6.192E-3:41:38:5:0:6:0

To reformat the vcf file I run

awk -F '\t' '/^#/ {print;next;} {OFS="\t";R=$4;n=split($5,a,/[,]/);s="";for(i=1;i<=n;i++) {s=sprintf("%s%s%s%s",s,(i==1?"":","),a[i],a[i]==R?"AAAAAAAAA":"");} $5=s; print;}' 001.vcf > 001new.vcf

Again run VariantsToTable on 001new.vcf But, still it shows the same error. I removed the particular line from vcf, error has changed to another line. The think which I noticed is REF column is same allele. here C,C Any help appreciated

Answer 1

[David Parry]

I don't think your awk code is doing what you intend. Below is a perl one-liner that should fix lines like the above where there are 2 identical ALT alleles:

perl -wane 'if (m/^#/){ print; next;} my @alts = split(",", $F[4]); if (@alts == 2 and $alts[0] eq $alts[1]){ $F[4] = $alts[0]; foreach my $gt (@F[9..$#F]){ my @formats = split(":", $gt); $formats[0] =~ s/2/1/g; $gt = join(":", @formats); } } print join("\t", @F) . "\n";'  001.vcf > 001new.vcf

However, this fix is a total hack and only applies to variants with 2 identical ALT alleles - it will not handle situations with more than 2 ALT alleles even if 2 or more of them are identical. The fact that you have lines like this in your input suggests something is going wrong when creating your VCF so I would strongly recommend re-examining how this is happening and if at all possible use a different tool to generate your VCF.

[DareDevil]

I tried this, still I get the same error