Multiple Psi-Blast Iterations From Command Line Using Remote Database

Question

[jdromano2]

When using PSI-BLAST from the command line, it does not allow you to specify successive iterations if the protein sequence database is on a remote server (i.e., using the '-remote' flag). I'm working on a bioruby script in which I need to iterate on PSI-BLAST until there are no new sequences added, but I need to use the remote nr database on NCBI's servers.

Is there a quick and easy way to get it to do these iterations for me without having to manually parse the results of PSI-BLAST and then using the results in a new PSI-BLAST call?

[rwipfle2]

Hey, did you ever figure this out? (Not even sure if you're still on this website anymore) I'm working on the exact same problem. Thanks.

[rosemont]

Hey, did you ever figure this out? (Not even sure if you're still on this website anymore) I'm working on the exact same problem. Thanks.

[nataliagru1]

Hello,

I am not sure if I am addressing your question however when you run PSI Blast from command line as such :

psiblast -query /dir/queryfile.fa -db /dir/example_database -evalue 0.001 -num_iterations 0 -outfmt 0 -out /dir/exp1_outfmt0.blast -out_pssm /dir/example.pssm -save_pssm_after_last_round

When you specify -num_iterations 0 PSI BLAST will iteratively search until convergence or until no new sequences are found.

[niamhlacyroberts]

How do you get the sequences from psi blast search on the command line? So e.g I've ran a psi blast search on a protein to convergence- how do I extract the homologs?

[nataliagru1]

Hello, I found this awk expression on biostars (cant find the original posting) to extract gene's from BLAST database based on blastp/blastn/psi-blast output format 6 and output to a fasta file.

awk '{OFS="\t"}{print $2,$9,$10}' /dir/sample.blast | xargs -n 3 sh -c 'blastdbcmd -db /dir/blast_database -entry "$0" -range "$1"-"$2" -outfmt %f >> /dir/output.fasta'

I hope this helps.