Any alternative for SSPACE-longreads or any suggestion to speed up

Question

[caizexi123]

Hi all,

I am using SSPACE-longreads to scaffold my genome (around 3G) with error corrected Pacbio data (10000000 reads), but the output is super slow, it seems the process will take 3 months. Is there any alternative to fulfill the same purpose? Or are there any suggestion to speed up, for example changing part of the code or any trick? BTW, I am using the default parameters.

[Sej Modha]

Did you try using more CPU by changing the -T parameter?

[caizexi123]

Yes, 8 threads. But, the blasr just take couple hours to complete. And I think SSPACE-longreads don't run multi-thread after blasr.

[Medhat]

what about PBJelly’s? Each step in PBJelly’s workflow can be run on a cluster

[caizexi123]

I am trying PBjelly now. Waiting for the result. But what else I can try?

[Medhat]

AHA
http://bioinfo-master.ird.fr:8080/smrtanalysis/doc/bioinformatics-tools/pbaha/doc/PbahaFunctionalSpecification.html

http://rhallpb.github.io/Applications/AHA.html

[caizexi123]

AHA is part of SMRT Analysis, it seems too complicate to setup. I will try after the PBJelly finish.

Answer 1

[felipe.zoujiro]

Maybe you can try OPERA-LG. It has a script for scaffolding using PacBio reads. Here is the paper https://genomebiology.biomedcentral.com/articles/10.1186/s13059-016-0951-y

It seems to be better than SSPACE-LongReads