Error, (in LoadSpliceMap) string to parse is too long

Question

[ziqing]

Hello, everyone I used OMA to find orthologs using 21 species genome. I exported the precomputed all-against-all for some species genomes in the oma database and then added my own species genome. First I get the result. But the second time I removed some species genome and add a new species genome, and then rerun, it encountered an error: Error, (in LoadSpliceMap) string to parse is too long.

These are the command I run:
Step1: ~/OMA242/OMA/bin/OMA -c /path/parameters.drw
Step2: qsub -t 1-100 -cwd job.sh
text in the job.sh: 
export NR_PROCESSES=100
~/OMA242/OMA/bin/OMA path/parameters.drw

Could someone give me some suggestions? Thanks in advance!

Answer 1

[Adrian Altenhoff]

Dear Ziqing

this error seems to be related to the splicing information file of one of the genomes. Given that it worked previously, I assume it must be from the newly added one. Does your genome have long fasta headers? For the problem to occur the lines in Cache/DB/*db containing the spliceing information need to be quite long, i.e. longer than 100k characters. The following shell commands should return you the length of all the lines in all database files. could you verify that there is at least one which is longer than 100000 characters?

 grep '<SPLICE' Cache/DB/*db | awk '{print length($0)}'

If this is the case, I wonder whether the spliceinformation you provided in the DB/*.splice make sense. If you think they are ok, I think we can find a way how to fix this.

Best wishes Adrian