Axiom PMRA cel files to VCF format

Question

[a.papadam]

Hello all, I am currently trying to convert CEL files from Axiom PMRA chip to VCF format. A VCF that will need to include also BAF and LRR values.

I am aware for the affy2vcf, but it did not work as expected.

To describe my exact issue:

My library files, dont work with apt-probeset-genotype ( the exact error is --> no more than one analysis is allowed ). I suppose problem with XLM file. I fixed the XLM file, keeping the thresholds. ) However, after affy2vcf many BAF values were higher than 1 or negative.

Any tips or ideas?

Answer 1

[Giulio Genovese]

I believe that apt-genotype-axiom can only generate TXT files while apt-probeset-genotype can generate both CHP and TXT files. But for the Affymetrix Axiom chips you might only be able to use the former. However, the TXT files can be used to generate a VCF file using bcftools +affy2vcf using a command like this:

bcftools +affy2vcf \
  --csv GenomeWideSNP_6.na35.annot.csv \
  --fasta-ref human_g1k_v37.fasta \
  --calls AxiomGT1.calls.txt \
  --confidences AxiomGT1.confidences.txt \
  --summary AxiomGT1.summary.txt \
  --snp AxiomGT1.snp-posteriors.txt \
  --output AxiomGT1.vcf

Just make sure you generate the calls/confidences/summary/snp files with apt-genotype-axiom. These will provide enough information to generate the BAF and LRR values. Do notice that bcftools +affy2vcf can generate, for homozygous calls, BAF values smaller than 0 and larger than 1 as the BAF values are not truncated. This can be useful as there usually is a residual correlation between BAF and LRR values that can be used to further adjust the values in downstream analyses