I am trying to use LEfSe to analyze differential abundance in some microbiome data. I’ve been using the microbeMarker package run_lefse command to run an analysis right off a phyloseq object, which is very helpful (link: Liner discriminant analysis (LDA) effect size (LEFSe) analysis — run_lefse • microbiomeMarker). However, I realized that there might be an issue with my data. I am compared two different groups with three samples each. However, I have biological replicates for each sample. I am worried that, if I just ignore the fact that there are replicates, I will be generating some misleading data because it looks like the input is 6 separate samples per group, while it is actually 3 samples with 2 reps each. Do you have any guidance for how to approach this to generate the most accurate results?
My initial thought was to just average the abundance of each OTU for the 2 replicates. I tried the “merge_samples” method in phyloseq to generate averages for each OTU for my 2 reps, but it doesn’t seem to be working properly and others appear to have issues with that command as well.
Thanks for the help!