ebi-pf-team / interproscan

Genome-scale protein function classification
Apache License 2.0
302 stars 67 forks source link

Fix hmmer SSI index failure for Superfamily and SFLD #350

Closed tgrego closed 8 months ago

tgrego commented 9 months ago

Superfamily and SSD original data files (HMM profiles) contain duplicated NAME / ACC values. This was not an issue in hmmer 3.1b2 but is in hmmer 3.3+. For that reason the 2 hmmer versions where in use by interproscan, but the problem would arise when using containers due to only a recent hmmer binary being used. The source data files where edited to in the case of Superfamily include virtual versions to the ACC (1234.5), and in the case on SFLD the NAME (name$2). Interproscan will ignore from the ./$ to maintain backwards consistency. Interproscan now can use hmmer 3.3 for all member databases.

https://www.ebi.ac.uk/panda/jira/browse/IBU-10729