One way to account for this (and which is commonly used) is to rarefy the samples: choosing a threshold to which you will subsample the reads at random to end up with the same number of reads per sample.

For example, if we compare sample A with 100,000 reads to sample B with 10,000 reads, the diversity in sample A will probably be greater just because we have a greater sampling depth. However, we need to account for variations in sequencing depth so I cannot use the raw counts here. I began with the OTU table that contains only samples above 1499 reads. I then compared the alpha diversity value between groups of samples. I chose to calculate the diversity within samples using two different metrics one phylogenetic metric (i.e.В takes into account the relatedness of the bacteria) and one non-phylogenetic metric (doesnвЂ™t take this into account). Now I move on to alpha diversity, which is a measure of the diversity within samples essentially we are asking, вЂњhow complex are these communities?вЂќ In the previous section, I explored the bacteria that were present in the samples and in the positive and negative sequencing controls. Case/control NPS differential abundance.Model 1 (case/control NPS including other covariates).Differential abundance with metagenomeSeqвЂ™s fitZIG.Correlating taxa in the sequencing controls.Cleaning up the OTU table and creating a phylogenetic tree.Reducing memory usage: dereplicating in parts.

R rarify samples