Closed #59 and opened this instead, I copied the text below from #59.
In working with the Northcott, et al. medulloblastoma data only (before including the Robinson, et al. cohort), I noticed that the "named" (e.g., associated with a pathway/gene set) LVs that had among the lowest FDR were associated with translation, RNA metabolism, etc. Here, I'm plotting two LVs (FDR < 0.05 in both datasets; subgroups comparison) related to translation. Medulloblastoma Group 4 has lower expression values for these latent variables. This is consistent with Forget, et al. Cancer Cell. 2018. wherein the authors observed that Group 4 had reduced pathway activity (using both mRNA and protein levels) for translation-related pathways. (They also noted that the protein levels of translation-related genes were higher than mRNA levels would suggest.)
Closed #59 and opened this instead, I copied the text below from #59.