Question about using WGS data

[katecycho]

Hello, Thank you for developing this great tool. I am wondering if I can use mutation calls from WGS. Thank you.

[bassanilab]

Hello some of the feature scores (such as ipMSDB score) can only be calculated for mutations in coding sequences. RNAseq expression will be predominantly 0 for WGS mutations. so the ML models presented here will not work well for WGS mutations. if you want to do a similar ML ranking for WGS mutation calls, then you have to remove the ipMSDB scores (and maybe other scores that don't work for WGS mutations) and retrain the classifiers. Also immunogenic mutations outside coding sequences are rare and its probably difficult to assemble a data set large enough to train a classifier.

Markus

[katecycho]

Thank you for your response. Would you recommend then to filter all coding regions from WGS and use that as input?