#9 Aborting process, probably due to an internal error.

[David Loyola]

Hello,
I ran latest version of mira using 14,456,692 iontorent reads for a EST/RNA assembly with those parameters:

job = denovo,est,accurate
parameters = --hirep_best COMMON_SETTINGS -GENERAL:not=24:keep_percent_memory_free=30 -SKIM:not=24
parameters = IONTOR_SETTINGS -AS:minimum_reads_per_contig=10

The assembly after 9 days failed, I have no idea because happened.

I look forward to your reply.
Best!

[Bastien Chevreux]

Hmmm ... the errors which are dumped out at the end make not much sense to me as they contain values which, in theory, are impossible. Also, the error appears in code which has been stable and ran error-free for a very long time now.

I am therefore asking myself whether this may not be a hardware problem. Can you please tell me whether
1) your computer has ECC RAM
2) whether the disk was somewhere near maximum capacity?

Generally, I see a couple of problems with your command line and the data set:
- you use "--hirep_*" with an RNASeq project although the documentation states that these switches are made for genome data. Please don't.
- your RNASeq data contains a quite large amount of incredibly repetitive data. If I had to guess: rRNA. This, together with the hirep_best switch you used, leads to temporary files reaching over 580 gigabyte as well as excessive run time

While I have no idea how the error you stumbled upon came into existence (though I suspect hardware), you should remove the hirep_best switch and then I suppose you'd get pretty good results in way less time.

Also, in case you want to, make a pre-assembly with a smaller subset (say, 1% of the data set and there maybe with the hirep_best) to find out which sequences are so highly overrepresented. Once you know, you can bait those out of the complete data set to make the life of the assembler easier.

Best,
B.

[Bastien Chevreux]

Closed, no further feedback was given.