Closed Rashesh7 closed 7 months ago
Hi,
Thanks for the interest in our package. Do you mean the simulation in the paper? The script used in the paper is at: https://github.com/friend1ws/pmsignature_paper/blob/master/analysis/simulation/script/runall.sh
Hope this helps for you.
Hi, Yes, the simulations used in the paper. Thank you for pointing me to the script. I had additional question, I am sorry if it is explicit somewhere, but when you mention True signature, are these from a specific file? or are you using the Cosmic Signatures as the base to simulate samples?
In this simulation, the signatures are generated from scratch by a probabilistic model (products of gamma distribution), and have no relationship with Cosmic Signatures.
Oh Ok, then does the tool have the capability of using an input set of signatures to make simulations?
No. We currently do not support that. For that experiment, you need to generate the set of mutations from e.g. COSMIC signatures with some probabilistic model by yourself.
Hi, Can you please elaborate on the Mutation Feature Vector Format for the type = "full" format
In the README you have described the format for type = "independent", right?
Basically, I have a file for 21 tumor samples (21 x 96) with number of mutations for each mutation context. Is there are way to directly use that kind of format?
Hi,
Can you please elaborate on how you generated the simulated data? For example, if you use K=3 while simulating the data, which 3 signatures are being used? And what are the ranges for alpha and gamma parameters. I am trying to compare a few tools including pmsignature and looking for a simulated data.