Blastp e-value, bit score, identify importance to identify homolog sequences

Question

[Mehmet]

Dear All;

I would like to ask you; which one or ones are more important to identify homolog sequences using blastp?

E-value?

Bit score?

Identify % ?

I am trying to find homolog sequences of candidate neuropeptides in my database by searching with blast p. I have an output (format 6). I want to filter results, then I will search in Uniprot o blast to identify my candidate proteins as neuropeptides or not. For this, which parameter should I use? Evalue, bit score, identify %, or combine them?

my approach will be to filter sequences which have e-value lower than 0.01 and bit score bigger than 90.

is this reasonable ? or is it possible to miss some neuropeptides.?

Thank you.

[berge2015]

Depends on if you want an identical hit or similar-enough peptides. Whenever I've had to do this kind of work, I've always put more emphasis on columns 4 (Alignment length), 5 (Number of mismatches), and 6 (Number of gap opens) first. Because, as long as alignment length is 100% with 0 mismatch AND 0 gaps, this alignment is the best one. Then you can move down the columns looking for evalue, identity %, etc.