Re: Extracting BLAST hits from a list of sequences

Given your datasets, perhaps the following will be helpful (it correctly returns the one BLAST entry which contains a matching fasta header):

use strict;
use warnings;

my %headers;

while (<>) {
    $headers{$1} = 1 if /^>(.+)/;
    last if eof;
}

local $/ = 'Query= ';

while (<>) {
    chomp;
    print $/. $_ if /(.+)/ and defined $headers{$1};
}
[download]

Usage: perl script.pl headerFile blastFile [>outFile]

The last, optional parameter directs output to a file.

The local $/ = 'Query= '; sets file reading to one BLAST entry ('record') at a time. The header info after Query= is captured and the complete entry is printed if that header is in the hash.

Comment on Re: Extracting BLAST hits from a list of sequences Select or Download Code

Replies are listed 'Best First'.
Re^2: Extracting BLAST hits from a list of sequences by no_slogan (Deacon) on Jan 20, 2014 at 16:44 UTC
Using `$/` is a good suggestion (and I upvoted it), but I feel that relying on the magic of the `<>` operator is unnecessarily clever in this case.	[reply]
Re^3: Extracting BLAST hits from a list of sequences by Kenosis (Priest) on Jan 20, 2014 at 18:32 UTC
Appreciate the upvote. It seemed only 'natural' to use `<>` in this case, since the OP already had files in `@ARGV` and using `<>` to process them avoided adding the code to do so.	[reply] [d/l] [select]