Sorry if this seems like a very basic question. Previously when I used cluster.split for a different dataset the summary file I got for my otus showed quite similar bacterial compositions of my samples to the summary file I got after classify.seqs with some small differences.
However, I have had to do a phylotype based analysis for a new study and the summary file outputted from that looks nothing like the summary from classify.seqs. There are far less unclassified sequences, which is good, but the quantities of bacterial genera, orders etc. seem completely different from their quantities as individual reads.
Is this normal? I’m not quite sure why binning things into taxonomies would cause this much change. Are taxonomies from phylotype analyses generally accurate and reliable?