How to obtain the SNVs belong to a typical mutational signature

[Liyang-Zhang]

Hi @FreekManders and everyone,

Is there a way to use MutationalPatterns to get the SNVs correspond to a specific mutational signature?

I assume that the absolute contribution means the number of mutations contribute to the signatures while the contribution values are noninteger. I am wondering how I could know which SNVs in my sample contribute to a specific mutational signature.

Thanks a lot!

Leon

[FreekManders]

Hi Leon,

Mutational signatures work by identifying mutation patterns in samples. When refitting signatures, the refitting algorithm selects signature contributions that best reconstruct the mutation profile of a sample. However, this does not determine which SNV belongs to which specific signature. Determining which SNV belongs to which signature is quite hard. One thing you can do is to see if the context of a mutation matches a specific signature, to roughly determine the odds it was contributed by that signature. A more accurate method is to train a ML model to predict if mutations belong to one specific signature, which was done here: https://doi.org/10.1016/j.stem.2021.07.012. This is however quite a bit of work and the accuracy will depend on how specific the signature you are training for is.

[Liyang-Zhang]

Thanks a lot @FreekManders for your advice. I will take a look at the paper.