I cannot really find a good answer to this question anywhere (eg. in the manuals of microbiota analysis software): Let say I have a study comparing the microbiota of two treatment groups. I have 5 replicates for each treatment group. Other than performing ANOVA type analyses where you can define treatment groups and replicates, what is the recommendation on treating replicates? I am primarily interested in generated PCoA plots and area plots, and having replicate variation encapsulated as one stacked bar, for example, would be better. Should the counts be averaged - that sounds like a bad idea, because library sizes can be different. Can the raw sequence reads (.fastq) for all replicates be supplied as a list for a single sample?
thanks for any suggestions and/or literature on this