Gene set enrichment analysis

[RunyuXia]

Hi! I have been trying to use DecipherSeq for analyzing my NMF results. However, I have encountered difficulties in annotating the themes of the modules. My modules seem to encompass a wide range of biological processes, with the main ones being generic modules. Is this normal, or have I done something wrong? Could you help me with this? I have attached examples of my module output below.
module1_bP_plot.pdf
module2_bP_plot.pdf
weighted_sub_B_cell.pdf
weighted_sub_CD4.pdf
weighted_sub_CD8.pdf
weighted_sub_myeloid.pdf
weighted_sub_NK.pdf

[lmurrow]

Without knowing more about the biology you're trying to study it is hard to say. DECIPHER-seq will help you identify gene programs that co-vary across samples in different cell types. So, it could be that the main variation across your samples is driven more "generic" processes like metabolism, etc. rather than meaningful changes in cell-cell signaling.

The main things to check are: 1) when you run iNMF, do not set batch to individual samples or if you do at least adjust lambda because DECIPHER-seq relies on finding programs that vary across samples (rather than programs that are shared across all samples). and 2) is there biology you expect to find, and do you see any gene programs that capture that biology?