I have a plant draft genome generated by ALLPATH-LG with NGS data. The genome size is ~1G. I am now using PBJelly2 to do scaffolding and gap filling with ~15X PacBio reads (average read length ~11Kb). The software is running smoothly without any error reported. However the assembly step seems like very slow and I estimate it will take 12 days according to the speed. I'm running the PBJelly2 on a single server with 56 CPU and 256Gb RAM. The command I used for the assembly step is "Jelly.py assembly Protocol.xml -x "--nproc=50" ". It still need 12 days... Is there any way to speed up?
Thanks,
Min
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
The --nproc parameter isn't going to help you in this situation. Instead, you need to fake that your multi-core machine is a cluster and set njobs in your Protocol to 50. As per the documentation Section V about creating your protocol.
If you have a single, large machine and not a cluster,
use the following command to submit all of your jobs
in the background and parallelize operations. Just be careful
the number of jobs/resources you execute or you can
freeze your system.
<command>${CMD}${JOBNAME} 2> ${STDERR} 1> ${STDOUT}&</command>
Hope this helps,
~/Adam English
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi Adam,
Thanks for this wonderful software.
I have a plant draft genome generated by ALLPATH-LG with NGS data. The genome size is ~1G. I am now using PBJelly2 to do scaffolding and gap filling with ~15X PacBio reads (average read length ~11Kb). The software is running smoothly without any error reported. However the assembly step seems like very slow and I estimate it will take 12 days according to the speed. I'm running the PBJelly2 on a single server with 56 CPU and 256Gb RAM. The command I used for the assembly step is "Jelly.py assembly Protocol.xml -x "--nproc=50" ". It still need 12 days... Is there any way to speed up?
Thanks,
Min
Hi Min,
The --nproc parameter isn't going to help you in this situation. Instead, you need to fake that your multi-core machine is a cluster and set njobs in your Protocol to 50. As per the documentation Section V about creating your protocol.
If you have a single, large machine and not a cluster, use the following command to submit all of your jobs in the background and parallelize operations. Just be careful the number of jobs/resources you execute or you can freeze your system. <command>${CMD} ${JOBNAME} 2> ${STDERR} 1> ${STDOUT} &</command>Hope this helps,
~/Adam English
Hi Adam,
I will try it next time.Thanks!
Min
You method works well. It runs much faster this time. Thanks!