Open Beigephage opened 4 months ago
You are somehow passing --msa-format-mode 0
to result2msa. Please don't use mode 0
.
Unfortunately I had not added a --msa-format-mode
flag at all, only mmseqs result2msa DB DB DB_clu DB_clu_msa
.
I just tried to run the command as mmseqs result2msa DB DB DB_clu DB_clu_msa --msa-format-mode 1
and I still got the same 3 files, not 1 consolidated msa.
--msa-format-mode 2
gave a file closer to what I was looking for, but some sequence headers seemed missing and upon inspection, those lines begin with ^@ when viewed via terminal. mode 4 gave a stockholm format. Is there a different option that provides the standard alignment?
trying to remove any hidden characters from the msa may not be solving the problem as it still cannot be used with reformat.pl
or addss.pl
(no secondary structure prediction after addss.pl)
I am still confused with what's going on. I can't reproduce the issue locally, everything works as expected.
Please provide the full command line calls, terminal output and directory listing of the folder with the result files. Maybe that will help me understand what's going on.
Running result2msa as described in the docs
mmseqs result2msa DB DB DB_clu DB_clu_msa
using older versions of mmseqs2 (in e.g 2019) previously merged results to give one '
DB_clu_msa
' file output with an msa. Newer versions of mmseqs2 (13.45111) running the same command produces multiple '.ffdata
' files, namely 'DB_clu_msa_sequence.ffdata
', 'DB_clu_msa_header.ffdata
', 'DB_clu_msa_ca3m.ffdata
' with no msa and no file similar to the old 'DB_clu_msa
' file.How can one retrieve the merged '
DB_clu_msa
' file in the old format?Thank you