I would like to run lefse on a dataset that is made up of all the single copy marker genes from SingleM. How do you recommend I do this to avoid errors with multiple hypothesis testing? Should I concatenate the results of each gene on each sample before I run Lefse or run lefse on the results of each gene and then combine the results?
Apologies for the delay in responding, I was away and am catching up on responses. My recommendation if you want robust multiple testing correction is to use our newer tool for differential abundance analysis, MaAsLin 2. It can accommodate a few different multiple testing corrections. LEfSe doesn’t have an explicit multiple testing correction; it filters by effect size to try to pull out likely candidate biomarkers.