MiBiG.gbk files cause BiG-SCAPE to crash
Also on behalf of Grimur Hjorleifsson:
Gene Cluster Family calling crashes.
Trying to run BiG-SCAPE on use MiBIG .gbk files (to tag the clusters generated by BiG-SCAPE) causes a crash.
To reproduce (from commit f00316e8):
Download MiBIG data from https://mibig.secondarymetabolites.org/mibig_gbk_1.3.tar.gz The ./mibig folder contains only the contents of the .tar.gz file.
Using Pfam-A version 31.0 (dated 24/02/2017) downloaded from ftp://ftp.ebi.ac.uk/pub/databases/Pfam/releases/Pfam31.0/ and processed according to instructions.
python ~/bigscape/bigscape.py -i ./mibig -o ./bigscape-mibig-regular --hybrids --pfam_dir ~/pfam --cutoffs 0.5
The program crashes with the following message:
Traceback (most recent call last): File "/home/grimur/bigscape/bigscape.py", line 2733, in htmlFolder=network_html_folder) File "/home/grimur/bigscape/bigscape.py", line 1651, in clusterJsonBatch a = domainGenes2allGenes[family][len(DomainCountGene[clusterNames[family]])-a-length] KeyError: 15
This might have been introduced by moving the Affinity Propagation logic to scikit-learn, as commit 792250d4 does not crash, given the same input.