Do you know if is that possible to identify bacterial enzymes (from microbiota) by analyzing transcriptomics data from a human tissue (i.e. TCGA samples)?
It may depend in how they capture the RNA for library prep, most of the experiments use poly-A capture (cheaper library) rather than rRNA depletion (Ribozero), and as RNA from bacteria lack of poly-A tail, It is not very likely to find this kind of data.