Dear deconstructSigs users,

I am using deconstructSigs on WES data from hypermutated tumors:

whichSignatures = whichSignatures(tumor.ref = sigs.input, signatures.ref = signatures.cosmic, sample.id = "MyPatient", associated = c(),signatures.limit = NA, signature.cutoff = 0, contexts.needed = TRUE, tri.counts.method = "exome2genome")

In input, I gave the number of mutations found for the 96 trinucleotides, without any normalization.

I hesitate between three values "default", "exome" and "exome2genome" for tri.counts.method parameter.

In the vignette description from August 29, 2016 (https://cran.r-project.org/web/packages/deconstructSigs/deconstructSigs.pdf), one can read in the description of whichSignatures(): " For exome data, the ’exome2genome’ method is appropriate for the signatures included in this package.

On github, one can read: "For exome data, the default method is appropriate for the signatures included in this package."

So default or exome2genome? Is it related only to the dataset included in the package?

My WES data were generated from V5 Clinical Research Exome capture kit, ~50% of the mutations are non coding (mainly in intron). Do you recommend to exclude the non coding mutations because of their trinucleotide repartition differences?

Thank you in advance for your precisions.

Dear,

I also have a similar question on that parameter. In my case we have WGS data so should we use 

tri.counts.method = "genome"
​instead of the "default"?

Thank you for your help,

Bastien