Is Music suitable for cancers with few somatic mutations

Question

[wangjiaqian]

Is Music suitable for cancers with few somatic mutations(usually <10 in whole exonic region)?

I think those types of cancer evolved with different mechanism of selection when compared to cancers with high mutation rate.

Answer 1

[Cyriac Kandoth]

Yes. MuSiC's SMG test was used to identify driver genes in Acute Myeloid Leukemia. See Fig 1b in the TCGA AML paper, where an average of only 13 mutations were found per exome.

MuSiC's SMG test makes no presumptions about your data, and is flexible enough to run on germline or non-human mutation lists. Per-gene mutation rates and background mutation rates (BMRs) are measured by the calc-bmr module, across the given cohort of cases. And the smg module tests the per-gene MRs against the BMRs to find the most significantly altered genes.

So as with any bioinformatics tool, you'll know whether it is suitable for your data, after you've run it through and taken a look at the results. Feel free to post a follow-up on Biostars.

[wangjiaqian]

Hi, Cyriac, the background mutation rate of every single gene is the same as overall bmr, is there anything wrong? All I do is to calculate gene coverage with a single bam for all sample, and followed running steps described here.

Thank you so much for your help!

[Cyriac Kandoth]

You're likely looking at the gene_mrs file. The last column just reports the overall BMRs that each gene's MRs are going to be tested against. This will not differ between genes unless you used the bmr_modifiers feature to correct for per-gene mutability bias (e.g. from transcription coupled repair). The per-gene mutation rate is not explicitly in that file, but the number of mutations and covered bps are in there. Num_Muts/Covd_Bps will give you per-gene categorized and overall MRs.

[wangjiaqian]

Hi, Cyriac, the background mutation rate of every single gene is the same as overall bmr, is there anything wrong? All I do is to calculate gene coverage with a single bam for all sample, and followed running steps described here.

Thank you so much for your help!