Applying A Pssm To Psi-Blast

Question

[white.d.andrew]

I'm able to generate a PSSM and a sequence from combinatorial experimental data. I want to BLAST the results to find proteins displaying these motifs. However, I have information about each position's sensitivity to substitutions, in addition to a consensus sequence. I know how to create a PSSM from this information, but I don't know how to incorporate this into a BLAST search. Does anyone know a cmdline argument to blast or a way of hacking a PSSM into a PSI-BLAST search to add position information to a sequence?

Answer 1

[Leonor Palmeira]

You might find some information in this post : http://www.biostars.org/post/show/1699/how-to-scale-a-pssm-for-ncbi-blast/

As well as here : http://etutorials.org/Misc/blast/Part+V+BLAST+Reference/Chapter+13.+NCBI-BLAST+Reference/13.8+blastpgp+Parameters+PSI-BLAST+and+PHI-BLAST/ and here http://gogarten.uconn.edu/MCB372/Laboratories/assign3.html

What you probably need to do:

• generate a .chk checkpoint file similar to the one generated by 'blastpgp' containing your PSSM

• feed it to blastpgp with something similar to:

  blastpgp -i test1.fa -d genome.faa -R test1.chk -o genome.test1.br -I T -a 2