Can we run VIPER (described here) with gene expression signatures generated from RNASeq-based differential expression experiments (e.g. DESeq, EdgeR, ...) instead of microarray data?

VIPER manual mentions "It can also take two matrixes as arguments, the first one containing the ‘test’ samples and the second the ‘reference’ samples"

If yes, which metrics to use (normalised read count, fold change, p-value...)?

How to pre-process this type of data?

Thanks !