We've done metaproteomics on a very simple enrichment culture (2 spp.) for differential expression. The data analysis basically shows functional changes in the community as a whole. Some reviewers have been suggesting us to partition the data and split the analysis between the 2 spp. in the system so we can evaluate the different responses of each organism without the bias introduced by changes in the community (there is an important change in the proportions of the 2 spp in the treatment vs control). Any takes on this? Would that really be appropriate?
PD: I've seen something called taxon-specific normalisation that it's supposed to remove this bias in metatranscriptomic data, any experience on that?
Could you isolate some pure cultures just like in beginning microbiology. Would that help you look at some interesting differences? Just a suggestion from a old chemist. Best wishes David Booth
Never split data. You cannot now split the data since it applies to the whole community, and interactions will still affect analysis of 2 species in isolation, since they grew in community. The only way to analyse 2 species at a time is to obtain data from them in isolation. Those reviewers are in essence saying it's okay to do individual t-tests on data that should be analysed by ANOVA, and using the individual results to comment on the whole community. It's a bad mistake and will send your paper straight to major revision with me, or rejection with an editor :)
- Badges
- Science topic