Open josemduarte opened 5 years ago
Thank you for reporting this. We estimate the sequence identity in default by the score per column. This is useful for lower identities. If you want the real sequence identity please use -a
.
See also https://github.com/soedinglab/MMseqs2/wiki#how-does-mmseqs2-compute-the-sequence-identity
Thanks @martin-steinegger ! Sorry I missed that in the docs. A possible usability improvement could be not outputting identity values in -a false
case.
Expected Behavior
A sequence search that match itself in target db should report 100% identity
Current Behavior
Identity reported is 91.4%:
Steps to Reproduce (for bugs)
My test_target_db.fa is:
and my test_query.fa is one of the sequences in the target:
The commands I ran are:
MMseqs Output (for bugs)
The .m8 output is:
Your Environment