I would like to BLASTP a list of protein sequences in fasta format against multiple protein databases. Since I'm only interested in the first hit of each database, it is possible that BLAST result would need to be parsed.

Do you know if this strategy is a good one, or you suggest me others?

Thanks for your help. Best regards, Bernardo

  • $\begingroup$ "Good" for what? If you're asking whether your approach with parsing is good, I'd say parsing should be performed for most of purposes any-case. If you want just one hit, there is a parameter 'max_target_seqs': see manual $\endgroup$
    – alephreish
    Commented Nov 30, 2013 at 23:00
  • $\begingroup$ We have identified vaccine candidates via an ex-vivo RNA-seq approach. Next step would be to perform conservation of these candidates (about 20) between multiple bacterial strains (about 200). I would like to know if it's a good idea to start with a BLASTP or go to more advanced strategies. $\endgroup$
    – biotech
    Commented Dec 1, 2013 at 12:32
  • $\begingroup$ Are you willing to set up standalone blast on your computer, and run a script on it? It's actually pretty straight-forward once you get to that level (using BioPython or BioPerl) $\endgroup$
    – adam.r
    Commented Dec 1, 2013 at 18:10
  • $\begingroup$ I do this all the time if you use stand alone blast. It's a simple flag you need to pass. $\endgroup$ Commented Jan 31, 2014 at 7:24

2 Answers 2


To scan multiple databases:-

  1. Download all the databases that you intend to match against (all databases will have a download option)
  2. Download standalone BLAST executables from NCBI
  3. Pool all the sequences from the different databases and make a BLAST database using the program makeblastdb (it is better to check for redundant sequences and remove them for improved performance)
  4. Run blastp with desired parameters against this database
  5. The first hit wont mean much. Use cutoffs on E-value or Bit-Scores or precentage identity

Now, if you vaccine targets a protein it is good to study the protein sequence conservation. BLAST isn't a really good option for studying conservation by comparing different orthologs. Use tools for multiple sequence alignment like ClustalW or T-COFFEE. In my opinion it is also better to use BLOSUM for scoring the alignments compared to PAM.

It might also be important to look at the protein structures. You can try VAST if possible.

  • $\begingroup$ Hi, maybe I didn't explained myself well. I already know that my vaccine targets a protein. My question is how to get yes/no table for 200 bacterial strains. ClustalW, T-COFFEE and VAST seem to be a latter step, but thanks for this. :) If I merge all the 200 bacterial genomes in the same database, would it be easier to parse later? that is, when checking presence/absence for each bacterial strain? $\endgroup$
    – biotech
    Commented Dec 4, 2013 at 16:50

If it's a species with already sequenced genome and you use a genome mapping approach, you probably get gene id's "automatically" during transcriptome assembly.

If it's a de novo approach, and all you get is a list of sequences, KAAS would help you to identify orthologous genes already known in other bacterial genomes (there you'll find even predefined subsets of genomes to blast against, 'for Prokaryotes' being one of them).

  • $\begingroup$ They bacterial genomes have not been published yet. I've been used a genome-based RNA-seq approach. $\endgroup$
    – biotech
    Commented Dec 4, 2013 at 16:42

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .