Dave Kristofferson writes:
" BLAST is fast without a doubt, but does it miss hits more
often than FASTA or is it of comparable sensitivity? I am sure that
one can pull up individual examples which can make either program look
more sensitive, but I'd like to hear the experiences of independent
parties that may have looked into this issue in some detail. I expect
that this question will come up frequently.
I have found BLAST to be useful in a slightly different context to
FastA or FastP.
It IS definitely fast. Also I am sure it can miss hits. I used both Fasta
and BLAST to test a novel fungal glutamate synthase....which I did not KNOW
was a glutamate synthase against the available databases.
The end result was that both FastA and BLAST came out with a suggestion that
the sequence I was using to probe with was a glutamte synthase (as both pulled
the glutamate synthases) but more importantly, BLAST allowed me to find
thedifferent regions of the function within the sequence.
I found the most use from using the peptide sequence. DIfferent "domains"
appeared that were common to very different proteins. Yet if a
memebrane spanning motif were present, BLAST finds this in disparate genes.
I put together all the regions that BLAST located from the novel protein and cam
e up with a likely function for the newly sequenced gene. Using both BLAST and
Fasta in the context of finding a probable function for an unknown gene
has thus been VERY useful.
I should point out that the synthases that were pulled by both FastA and BLAST
were prokaryotic and *VERY* distant to the fungal one I was using to test.