I have found that the number of sequences reported to be removed does not always match the number that have in fact been removed. It seems that remove.seqs simply reports the number of records in the input *.accnos file, rather than actually comparing the # of records in the *.fasta file before vs. after the command is run. If there is an accession number in the *.accnos file that does not match any sequences in the .fasta file, the reported # of sequences removed will be inaccurate. Alternately, if I try and use the output of align.seqs (.flip.accnos) with remove.seqs, I sometimes get a report of a higher number of sequences being removed than there are accession #s in the *.flip.accnos file.
Thanks for bringing this to our attention. Mothur version 1.21.0 will include accurate counts for each output file.
I’ve noticed what looks like the same problem with the groups file – remove.seqs does not remove anything from it. I gave remove.seqs an accnos file and a groups file, and a *.pick.groups file was generated, but it is identical to the *.groups files given to remove.seqs.
Has this been fixed in v1.21? I’m in the middle of an analysis that I’d like to finish before I update…
Thanks, Lee Ann