BLAST+ and blasting against the NCBI database

[kevluv93]

Hey guys, I'm trying to run blastx on a transcriptome against NCBI's database so I can get an annotation. I'm using a HPC with trinotate on it, and I'm assuming that I've installed blast+ correctly.

How do I BLAST against the NCBI website's database? I'm aware that you can configure any database file for blast+ to run data against, but what if I want to run a large amount of RNA contigs (like 60,000) against NCBI's website database? I'm aware of the -remote option, but this option will default my computer to using one core to run this job. I assume that the -remote option will let my data run on NCBI's HPCs, but I don't want my job to get timed out.

Any help? How do you approach using blastx on a large dataset?

Thanks!

[GenoMax]

For that large number of input sequences you would want to run this on a local cluster by splitting your input into multiple file and then running the jobs in parallel. I don't think the "-remote" option is meant for huge datasets.

You may want consider against swissprot/trembl or refseq databases to limit your search space. Is this search for annotating a new transcriptome?

[kevluv93]

Yes, it's a new transcriptome for copidosoma. I've tried running this data against uniprot, but I keep on getting low quality hits for drosophila and humans. Of course, I don't know what good outputs for this animal would look like. I'd assume I'd get good hits for Nasonia, assuming that the uniprot database has Nasonia vitripennis in it.

So,in essence, I can either divide my file and run -remote or use an existing database from an animal like Nasonia to get my annotations?

[GenoMax]

If Nasonia is the closest relative you can use then search against this Ensembl protein set: ftp://ftp.ensemblgenomes.org/pub/met...tripennis/pep/ or from Nasoniabase: http://hymenopteragenome.org/nasonia...v1.2_pep.fa.gz Annotation appears to be there too: http://hymenopteragenome.org/nasonia...GSv1.2.gff2.gz

The -remote option is probably not meant for 60K items. NCBI may ban your IP if you try to launch too many jobs.

[AntonioRFranco]

Take a look to Blast2Go
The pro version will speed your searching by using their Cloud faciltites, and you can use your own local databases as well