I am wondering about the proper way to control for gene-specific differences in the background mutation rate when using Genome MuSiC PathScan.
From what I understand from the documentation, PathScan assumes a uniform background mutation rate for all genes that enter the analysis. I think this assumption is wrong and I am thus worried about its consequences in terms of pathways that are erroneously identified as significantly mutated.
Is there any way to deal with this? Is it for example a viable strategy to first identify significantly mutated genes with Genome MuSiC SMG (which considers gene-specific differences in the background mutation rate) and let only the significantly mutated genes enter the pathway analysis? But I assume that this would result in a loss of power, because we are no longer looking at all mutations affecting a particular pathway at once.
Are there maybe alternative tools that consider gene-specific mutation rates for pathway enrichment analysis in cancer?
Good question. The assumption of uniform background rates is completely inappropriate. I'm trying to deal with the same issue using hierarchical models, but it's a bit tricky; so I'd be very interested to know of any other tools