How does cluster.split work without a taxonomy? reads the dist then picks a grouping level? I’ve tried using cluster.split with the tax file but still run into resource limits because one dist, I’d guess the proteos, is still a huge file. Would you recommend choosing a lower tax level (I was using 3) or cluster.split without the tax?
Without the taxonomy it looks at the distance matrix and figures out how to split it apart so that there aren’t small distances that span the submatrices. I’d suggest going to level 4 or 5 (order and family).