fusioncatcher
fusioncatcher copied to clipboard
ndaniel
fusioncatcher-build fails for canis lupus familiaris, with IndexError: list index out of range
This issue is similar to #82 from 2018, except it could not be solved by changing servers, so the root cause might be something different this time. The latest version of FusionCatcher was installed with the recommended method. The error message is the following (the full log file is attached):
Traceback (most recent call last):
File "/home/joakim/bin/fusioncatcher_24_02_2021/fusioncatcher/bin/add_custom_gene.py", line 288, in
I am trying to reproduce the bug and let's see. At first glance it looks like Ensembl has changed the organism name from canis_familiaris to canis_lupus_familiaris.
Yes, they have changed the name. The script does download some files, but some of them, such as exons.txt end up as empty files. On line 285-288 in add_custom_gene.py, the script then tries to read from this file (exons.txt) and gets nothing, resulting in the error, I think.
Yes, indeed I can reproduce the bug and it is related to the change from canis_familiaris to canis_lupus_familiaris in Ensembl. Several scripts need to be modified. Soon I will push the changes here in Github but I will not release a new official version of FusionCatcher yet.
Shortly, these two lines
ense = options.organism.lower().split('_',1)
ensembl_organism = ense[0][0]+ense[1]+'_gene_ensembl'
should be replaced with these two lines
ense = options.organism.lower().split('_')
ensembl_organism = ense[0][0] + ense[1] + '_gene_ensembl' if len(ense) == 2 else ense[0][0] + ense[1][0] + ense[2] + '_gene_ensembl'
in the following files:
- get_biotypes.py
- get_exons_positions.py
- get_genes_descriptions.py
- get_hla2.py
- get_mtrna.py
- get_paralogs.py
- get_refseq_ensembl.py
- get_rrna.py
- get_trna.py
In get_paralogs.py, it seems the following line needs to be changed as well: #org = ense[0][0] + ense[1] org = ense[0][0] + ense[1] if len(ense) == 2 else ense[0][0] + ense[1][0] + ense[2]
@jowkar Indeed, that is correct!
It looks like after these fixes there are still more things to fix.