Open torognes opened 7 years ago
Dr. Rognes and the VSEARCH team,
Do you have any update about this ticket? In particular, there is a bug in Clustering where the consensus sequence given for a cluster can sometimes be incorrectly trimmed in version 2.14.2. Has this bug been fixed in version 2.21 yet? Thanks!
I don't have any updates on this issue, unfortunately.
I was not aware that there was a bug related to consensus sequences in version 2.14.2. Could you give some more details? Do you have an example that shows the problem?
So, the methods for doing multiple sequence alignments and consensus sequences have not been changed since version 2.14.2?
I am not sure if I am able to find an example for this bug, since it sometimes happened, but not always. I can try to find it, and let you know. Meanwhile, may I suggest that you to test and compare the consensus sequences using different datasets with those from USEARCH/Clustering? This way, you may find the issues quickly on your own.
The multiple sequence alignment algorithm used in VSEARCH, the centre-star method, is rather simple. Consider implementing a better algorithm.
As a start, unaligned residues to the left of the first residue in the centre sequence could be right-aligned instead of left-aligned.