How to control for UV mutational signature? (for recurrent somatic mutation)

Question

[ConvolutedGenome]

Hi,

I have a list of 100,000+ non-coding regions and I want to find noncoding regions that are recurrently mutated. Now my study is in melanoma, and melanoma tumours are known to be highly mutated at UV signature sites.

This mean that I HAVE to consider this as a potential covariates. The problem is that I don't know how :(

Here's what I'm thinking:

• Count how many UV signature sites I've got in each of these regions.
• Use this UV count as a sort of "feature" for clustering. My hope is that regions with similar UV site counts might behave similarly mutation-wise.

But here's where I'm stuck:

• How do I even find these UV sites? Is there a BED file or something out there that can help me figure out where these UV sites are?

Any help is much appreciated. Thanks you so much!

Answer 1

[Sleepy]

Hi,

I'm facing exactly the same problem, did you figure it out at last?

BTW, I found the methods of this article (doi: 10.1038/s41586-020-1965-x) mentioned the possible solution.

They said: "For all tumor cohorts, we excluded elements with more than 50% of mutations attributed to APOBEC mutation signatures (BI_COMPOSITE_SBS signatures 2_P , 13_P and 69_P)".

It seems that they firstly assigned each mutation a mutation signature and then excluded elements with more than 50% of mutations attributed to some mutation signature class. But I don't know how to assigned a mutation signature to a single mutation and even whether it's possible to do that.

Thanks!