Hi,
thank you very much for a great tool, which has proven very useful for our analyses!
I noticed that I got different results when running mmsig in the 2 modes, "vcf" and "classes" on the same samples. Upon looking into this in more detail, I figured that the reason was probably that the trinucleotide order returned by the function mut.to.sigs.input() (imported from deconstructSigs) returns the trinucleotides in a different order than my reference signature file. I solved this for myself by adjusting the order manually. However, it might be useful to include a consistency check in the code as well.
Many thanks, Verena