Are There Any Caveats When Creating A Maf File For Music?

Question

[n]

Hi all,

The current version of MAF is 2.4. Does MuSiC support this or has any issues with this version, if we make a mutation file that conforms to its specification?

Are all of the MAF columns required? If not, which ones can be excluded and is the ordering of the columns important?

Many thanks in advance.

Answer 1

[Cyriac Kandoth]

Nope, to both questions.

The changes from MAF v2.3 to v2.4 are detailed here. Of the columns that MuSiC cares about, only Variant_Classification has some changes - the values De_novo_Start_InFrame and De_novo_Start_OutOfFrame are no longer allowed. But MuSiC still treats them as valid values for backward compatibility (see this line of code).

The order of the first 34 MAF columns is important. But only the following columns are actually used by the tools in MuSiC, while the others can be left blank, or can contain gibberish. Hugo_Symbol, Chromosome, Start_position, End_position, Variant_Classification, Variant_Type, Reference_Allele, Tumor_Seq_Allele1, Tumor_Seq_Allele2, Tumor_Sample_Barcode

Some tools in MuSiC require additional non-standard columns in the MAF (ordering doesn't matter, as long as they're after column 34). For example, the tools cosmic and proximity need the MAF to specify the transcript_name (doesn't matter what convention as long as its consistent), amino_acid_change(protein change in HGVS), and c_position(codon change in HGVS) that each variant was annotated to.

[Marina Manrique]

Hi Cyriac,

I've been using your script vcf2maf to get the MAF for working in Music. I could annotate it with VEP easily.

My question is, how do I add the columns transcript_name, amino_acid_change and c_position to it?

Thanks a lot!