ebi-pf-team / interproscan

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

Issue with superfamily via conda install #248

Closed AdamStuckert closed 2 years ago

AdamStuckert commented 2 years ago

Hi,

I've installed version 5.54-87.0 via conda. I followed the instructions to download the tarball and move the data files into the appropriate location. But, when I try to run the initial setup I am getting errors with the superfamily library that comes with the install.

Checking any hmm models that need indexing ... this may take a few minutes
Working...    SSI index construction failed:
  primary keys not unique: '7,8-didemethyl-8-hydroxy-5-de' occurs more than once
Working...    SSI index construction failed:
  secondary keys not unique: '100879' occurs more than once
Completed indexing the hmm models.

Then, when I try the initial test run with ./interproscan.sh -i test_all_appl.fasta -f tsv -dp I get this error:

Error: Failed to open binary auxfiles for data/superfamily/1.75/hmmlib_1.75: use hmmpress first

Running hmmpress data/superfamily/1.75/hmmlib_1.75 does not fix the issue, and yields the same error about secondary keys not being unique.

Head of the superfamily library:

HMMER3/b [3.0 | March 2010]
NAME  0034782
ACC   54076
DESC  RNase A-like
LENG  124
ALPH  amino
RF    no
CS    no
MAP   no
COM 1 This model under the license at http://supfam.org/SUPERFAMILY/license.html

Thanks, Adam

matthiasblum commented 2 years ago

Hi,

This issue has already been reported (see #232).

To avoid this error, you can use an older version of HMMER (3.1b2) to index the profile databases with duplicated keys (SUPERFAMILY and SFLD).

Lsz-20 commented 2 years ago

@matthiasblum I have also installed 5.54-87.0 via conda,the hmmer=3.3.2 & hmmer2=2.3.2.Then i change to hmmer=3.1b ,but the results have not change...