Hello,
I was little bit confused regarding the validation dataset files. I have developed different model for different studies (EMTAB, HOVON, UAMS) since the number of microarray expression values in these studies are different. In the validation dataset, the study names are different for e. M2gen, DFCI, Hose. Can you please that study in validation is corresponding to which study in training data ?? Or how to deal with the validation dataset ??
Created by Piyush Agrawal piyush Dear Piyush,
The validation studies have a mix of microarray and RNASeq as indicated by the clinical annotation files and the wiki. The validation studies do not correspond to the training studies in any one to one manner but a completely independent studies. A model may need to normalized datasets and/or summarize gene/probe signatures and handle missing gene/clinical data.