chg60
commented
3 weeks ago Hello!
Thanks for reporting this unexpected behavior. I have only ever seen that error with MMseqs2 (including the most recent version) when clustering groups of very similar nucleotide sequences, so I would be interested in tracking down the source of the issue on your protein set, provided the sequences aren't proprietary in some way that prevents you sharing them!
If you can share the sequences, please attach them in a reply to this message and provide the command you used to run PhaMMseqs when you got the error. If possible, it would also be helpful to know your CPU/RAM and operating system.
I have been in a different lab for almost two years, so I have somewhat limited time to address this issue. However, with that said, I have been actively working on a version of PhaMMseqs that can use any of the recent releases of MMseqs2 (13.45111, 14.7e284, and 15.6f452) that should be ready for public use sometime in the next month. I also like the idea you posed of allowing users to supply their own clustering output and simply performing the pham alignments and pangenome analyses from there. I'll try to incorporate this feature in the next release also!
I look forward to hearing from you soon with more details!
Best, Christian
Hi developer,
Thank for developing the phammseqs!
I noticed that only mmseqs2=13.45111 is required for running phammseqs. I am encountered error associated with mmseqs2 during the clustering of protein sequences (there must be an error: xx deleted from xx that now is empty, but not assigned to a cluster), and maybe this error due to old version of the mmseqs2.
Can we used more recent version of mmseqs2 in phammseqs? or can we use our own clustering output from mmseq2 as input for phammseqs for the downstream analysis.
Thanks!