Searching Homology (Blast)

Question

[onpelikan]

Hi, I need to analyze protein mutations via MAPP tool. For this tool I need have multiple alignemnt file and phylogenetic tree file. So I need to obtain homology sequences, but not so closely related ones, because closely related sequences are not significat for MAPP. How to use BLAST to find not so close related homolgies. I tried use PSI-BLAST but:

1. How many iteration is needed?
2. Is inclusion_ethresh param significant?

[Josh Herr]

There are many ways to measure protein mutations, so I'm unclear why you need to use MAPP. Of course you'll need to have a curated alignment to assume homology, but are you asking how to get a phylogenetically diverse array of sequences in a given lineage?

Answer 1

[Naren]

Blast will not give you alignment file and tree file as output.
Clustalw2 executable file can help you.
Options are available for different outputs of alignment and tree files.

MAPP requires
1.fasta alignment file which matches with alignment.fas obtained from clustalw2
and
2.tree file which matches with tree.dnd obtained from clustalw2

Use these files for running MAPP:
Here is your code. for windows cmd:

java -jar MAPP.jar -f alignment.fas -t tree.dnd -o out.txt

ENJOY!!!

[onpelikan]

Thank you for your answer but my question is about psi-blast. How to obtain not so closely related sequences? For multiple sequence alignment I'm using Clustal Omega and for tree I'm using SEMPHY (it is not very well. Probably MEGA or PhyML would be better for tree reconstruction).