Alt allele - after bcftools norm

Question

[kamanovae]

Hi!

I created a vcf file based on rsId and information from dbsnp. Then was ran the bcftools norm to make the file suitable for GATK.

bcftools norm -f ref/unmasked_ref.fa --check-ref w -m - rs_coord.hear_color_with_rs.vcf

But I still have "-" in the ALT column. I can't start indexing the file and I think that GATK will not process such lines.

How can I cover such lines from this:

chr22   42127856        rs5030867       T       G       .       .       RS=rs5030867
chr22   42128173        rs5030656       C       CCTT    .       .       RS=rs5030656
chr22   42128176        rs5030656       TCT     -       .       .       RS=rs5030656
chr22   42128242        rs35742686      T       -       .       .       RS=rs35742686
chr22   42128945        rs3892097       C       A       .       .       RS=rs3892097
chr22   42128945        rs3892097       C       G       .       .       RS=rs3892097
chr22   42128945        rs3892097       C       T       .       .       RS=rs3892097
chr22   42129084        rs5030655       A       -       .       .       RS=rs5030655
chr22   42130692        rs1065852       G       A       .       .       RS=rs1065852
chr22   42130692        rs1065852       G       C       .       .       RS=rs1065852

To this (add nucleotide from reference before alt allel):

chr22   42129084        rs5030655       TA       T       .       .       RS=rs5030655

[iraun]

Have you tried --check-ref s ? From bcftools manual, it looks like w just reports a warning, without attempting to fix the reference allele as s does.

[kamanovae]

Unfortunately this approach doesn't work, because the REF alleles are fine. I didn't find a flag that could fix ALT

[iraun]

Right, I misread the question, sorry about that :). FYI, It seems that the issue was somehow addressed in this stackexchange post, have you tried the solutions suggested there?