Reciprocal blast (RBH) - help

ac91 · 08-23-2015, 04:02 AM

Hi!

I´m new in the bioinformatic world so be patience if my question is dummy.

I have two fasta files and I run blast reciprocal best hits (RBH) for blastp using the script from peterjc that you can be found here: https://github.com/peterjc/galaxy_bl...ools/blast_rbh

So, the output gives me:A_id, B_id, A_length, B_length, A_qcovhsp, B_qcovhsp, length, pident and bitscore.

But I´m also interested in other options as e-values etc...

Anyone knows if there is an away of change the script that I´m using to obtain other results or if there is another script that can give these options ?

thanks!!

maubp · 08-23-2015, 05:19 AM

I'm glad it's useful (our preprint covering this and the Galaxy BLAST+ wrappers is out, and the formal publication in GigaScience is coming soon - so you'll be able to formally cite the script).

The e-values depend on the size of the FASTA files (or the temporary databases made from them), so are often not directly useful for comparison between datasets - that's why I deliberately only reported the bit score.

But yes, it ought to be fairly simple for you to add the evalue field to the BLAST commands and collect that in the final output. Do you know any Python programming?

ac91 · 08-23-2015, 06:11 AM

First of all, thanks for your fast answer.

I had the idea that e-values were the main criteria to "eliminate" some hits. I´m wrong?

Yes, I have some basic knowledge in python programing.

maubp · 08-23-2015, 01:25 PM

Maybe read http://dx.doi.org/10.1371/journal.pcbi.1000160 (linked to from the tool help in the Galaxy wrapper XML file).

My RBH tool does not use the e-value at all, the "best" hit has the highest bit score (which in turn means it has the lowest e-value because BLAST calculates that from the bit score).

ac91 · 08-23-2015, 01:34 PM

Thank you very much!!!

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
reciprocal blast	renesh	Bioinformatics	7	02-10-2017 09:33 AM
Reciprocal Blast Guidance...	James88	Bioinformatics	4	11-05-2014 03:45 AM
Reciprocal blast help	Shorash	Bioinformatics	5	07-21-2014 11:39 PM
detect orthologs using reciprocal BLAST	nicedad	Bioinformatics	2	06-07-2012 08:37 AM
Need Help Regarding Reciprocal best hit in BLAST	Bharat	Bioinformatics	1	03-11-2010 04:04 AM

08-23-2015, 04:02 AM	#1
ac91 Member Location: porto Join Date: Aug 2015 Posts: 13	Reciprocal blast (RBH) - help Hi! I´m new in the bioinformatic world so be patience if my question is dummy. I have two fasta files and I run blast reciprocal best hits (RBH) for blastp using the script from peterjc that you can be found here: https://github.com/peterjc/galaxy_bl...ools/blast_rbh So, the output gives me:A_id, B_id, A_length, B_length, A_qcovhsp, B_qcovhsp, length, pident and bitscore. But I´m also interested in other options as e-values etc... Anyone knows if there is an away of change the script that I´m using to obtain other results or if there is another script that can give these options ? thanks!!

08-23-2015, 05:19 AM	#2
maubp Peter (Biopython etc) Location: Dundee, Scotland, UK Join Date: Jul 2009 Posts: 1,543	I'm glad it's useful (our preprint covering this and the Galaxy BLAST+ wrappers is out, and the formal publication in GigaScience is coming soon - so you'll be able to formally cite the script). The e-values depend on the size of the FASTA files (or the temporary databases made from them), so are often not directly useful for comparison between datasets - that's why I deliberately only reported the bit score. But yes, it ought to be fairly simple for you to add the evalue field to the BLAST commands and collect that in the final output. Do you know any Python programming?

08-23-2015, 06:11 AM	#3
ac91 Member Location: porto Join Date: Aug 2015 Posts: 13	First of all, thanks for your fast answer. I had the idea that e-values were the main criteria to "eliminate" some hits. I´m wrong? Yes, I have some basic knowledge in python programing.

08-23-2015, 01:25 PM	#4
maubp Peter (Biopython etc) Location: Dundee, Scotland, UK Join Date: Jul 2009 Posts: 1,543	Maybe read http://dx.doi.org/10.1371/journal.pcbi.1000160 (linked to from the tool help in the Galaxy wrapper XML file). My RBH tool does not use the e-value at all, the "best" hit has the highest bit score (which in turn means it has the lowest e-value because BLAST calculates that from the bit score).

08-23-2015, 01:34 PM	#5
ac91 Member Location: porto Join Date: Aug 2015 Posts: 13	Thank you very much!!!