Currently, I’m trying to incorporate the HumanN3 and MTX models into my work, but I’m encountering challenges when it comes to combining MTX and MGX data.
I have two different enriched bacterial cultures from the same stool sample. While I have MTX data from both cultures, I only have one paired MGX dataset, which includes MGX from one of the enriched cultures as well as MGX from the stool.
To compare the differential gene expression between these two enriched communities, I’m unsure whether it would be better to combine both MTX data with the MGX from the stool, or to rely on the taxonomic profile built on MTX.
If I’m following this correctly, it seems reasonable to me to use a single taxonomic profile from the one MGX sample to decide which species to map the various MTX samples against (since they’re all based on the same source sample). In other words, pass the one MGX taxonomic profile to HUMAnN via the --taxonomic-profile flag.
I would NOT use that one profile to normalize all of the MTX samples though, as it’s possible that some bugs increased/decreased in DNA abundance after making aliquots from the original sample. An alternative approach would be to normalize the MTX data within each bug from each MTX sample. That still has the effect of normalizing MTX for DNA copy number, it’s just not quite as good at differentiating RNA 0s due to missing genes from other factors.