skoren
commented
6 years ago - Whatever jobs are completed won't be run again. Just run the original canu command.
- The correctedErrorRate won't affect the read correction. Your rate is set to 0.065 which is not that high, I'd actually raise it even more to 0.105 if this is Sequel data which is lower quality than RSII.
- You don't need to manually adjust the time, just tell canu to request more for those jobs (or all jobs if it's easier). Add
gridOptions="--time=168:00:00which will make all jobs ask for 7 days. You can also dogridOptionsCorMhap="--time=168:00:00to change just this step.
zrlewis
Hi again,
I am working on assembling a 3.3 Gb invertebrate genome using about 50X PacBio coverage (160 Gb reads). With your suggestions for specifying grid options (Issue #886) I was able to get the assembly running properly on a SLURM partition with many available nodes.
Here is my submission script:
I will paste my
canu.outlog below.I based my
gridOptionscormhaptime setting on previous runs of this same assembly that had taken only a few hours to complete, but now the jobs are starting to timeout. This step has been running for about two weeks so far and only about 25% of the jobs successfully completed, with the rest timing out eventually. I can have the times manually adjusted on our cluster to get some to complete, but I have a couple questions:cormhapjobs, willCanupick up where it has left off? Would I just re-submit the original submission script? This is related to Issue #794.correctedErrorRatehigh, because this is a wild-caught individual with probable high heterozygosity. Perhaps it is too high and this is bogging down correction. Could this be an issue? If I stopCanuand resubmit with a lower setting, will it create problems?cormhap?Thanks in advance for any suggestions!
Here is
canu.out: