vcf2maf v1.2.1

• Hotfix for snpEff/VEP prerun checks
• Support VarScan VCFs that store reference allele counts separately

vcf2maf v1.2

• Now parses out tumor/normal allele read counts from VCF to print in MAF
• Handles multi-allelic variant sites by choosing allele with highest read support
• Support for complex indels that include substitutions
• New test VCFs for MuTect and VarScan

Note: VCFs from VarScan use hardcoded sample IDs TUMOR and NORMAL in the genotype columns, that likely differ from the real sample barcodes that you want to report in the MAF using --tumor-id and --normal-id. This caveat can now be handled with --vcf-tumor-id TUMOR --vcf-normal-id NORMAL.

vcf2maf v1.1

• Now defaults to Ensembl's VEP for better HGVS format, and canonical transcript selection
• Now accepts VCFs that are already annotated with VEP or snpEff
• Supports user-specified directories for VEP/snpEff installation and DB cache
• Simplified code that prioritizes and selects transcripts to report in the MAF

For each variant, the protocol for selecting a single affected transcript is now as follows:

1. Sort effects first by transcript biotype priority, then by effect severity, and finally by transcript length
2. Pick the gene affected on the top of the list, and choose it's canonical transcript (VEP only)
3. If the gene has no canonical transcript tagged, choose its longest transcript instead

vcf2maf v1.0

Finds effects of each variant on all possible transcripts using snpEff, and selects the "most deleterious effect on the best transcript". If there are multiple such affected transcripts, it chooses the longest one.

Future releases will add different ways to select a single effect and transcript per variant.