Multiple reference alleles in vcf file

Question

[waqaskhokhar999]

I am interested to petform splicing QTL analysis (sQTL). In my vcf files at some reference positions, I have more than one allele, should I need to keep them or remove rows containing those snps? For example position 187 and position 194 contains more than one allele so should I need to remove these rows?

#CHROM  POS ID  REF ALT QUAL    FILTER  INFO    FORMAT  108 139
1   73  .   C   A   .   PASS    .   GT  0   0
1   83  .   T   C,A .   PASS    .   GT  1   1
1   187 .   TG  T   .   PASS    .   GT  1   1
1   188 .   G   T   .   PASS    .   GT  0   0
1   189 .   T   C,G .   PASS    .   GT  0   0
1   190 .   G   A   .   PASS    .   GT  0   0
1   194 .   ATT A   .   PASS    .   GT  1   1
1   209 .   C   T   .   PASS    .   GT  0   0

[Ram]

I don't see anywhere that you have multiple REF alleles. There are multi-allelic sites (with multiple ALT alleles), sure, but no multiple REF alleles. Maybe you're looking at the wrong column header? Here's your data formatted for eyeballing:

#CHROM  POS  ID  REF  ALT  QUAL  FILTER  INFO  FORMAT  108  139
1       73   .   C    A    .     PASS    .     GT      0    0
1       83   .   T    C,A  .     PASS    .     GT      1    1
1       187  .   TG   T    .     PASS    .     GT      1    1
1       188  .   G    T    .     PASS    .     GT      0    0
1       189  .   T    C,G  .     PASS    .     GT      0    0
1       190  .   G    A    .     PASS    .     GT      0    0
1       194  .   ATT  A    .     PASS    .     GT      1    1
1       209  .   C    T    .     PASS    .     GT      0    0

[waqaskhokhar999]

Indeed this is a new thing for me, I have again checked the original file and it contains multiple reference alleles, I have downloaded the vcf file from here.

[Ram]

Can you please paste a few sample lines? Use this line of code to get the sample records:

awk -F"\t" -v OFS="\t" -v cntr=0 '$4 ~ /,/ { cntr=cntr+1; print; } cntr==10{ exit; }' | column -ts $'\t' vcf_file.vcf

Answer 1

[Fabio Marroni]

Positions 187 and 194 are deletions. So in your reference you have TG and in the alternative allele you have T (deletion of a G) . Same is true for position 194, were ATT is the reference allele and the alternative allele is A (meaning that TT are deleted).

[Ram]

Neither of those positions fit the description of "multiple alleles". They are both single multi-base alleles, the standard way of denoting a deletion.

[WouterDeCoster]

But that notation might explain OPs confusion.

[Ram]

That makes sense. Just noticed OP referring to these positions specifically, so they should probably read the VCF specification and make sure they understand regular representation versus multi-allelics..