From bf85a36724ee52470bac86c598b03c2b90132658 Mon Sep 17 00:00:00 2001
From: elserj <elserj@localhost>
Date: Fri, 17 Sep 2010 23:09:10 +0000
Subject: [PATCH] Fixes the gene headers in the fasta files so that they are
 correctly parsed by blast.  Changes it so the only info in the header is the
 protein/gene id.

svn path=/; revision=37
---
 interactome_scripts/fasta_header_fix.pl | 35 +++++++++++++++++++++++++
 1 file changed, 35 insertions(+)
 create mode 100755 interactome_scripts/fasta_header_fix.pl

diff --git a/interactome_scripts/fasta_header_fix.pl b/interactome_scripts/fasta_header_fix.pl
new file mode 100755
index 0000000..eaaa1a8
--- /dev/null
+++ b/interactome_scripts/fasta_header_fix.pl
@@ -0,0 +1,35 @@
+#!/usr/bin/perl
+
+use strict;
+use warnings;
+
+require "$ENV{HOME}/scripts/jaiswallab/interactome_scripts/find_species.pl";
+
+my @files;
+
+if($#ARGV == 0) {
+	@files = $ARGV[0];
+}else{
+	@files = glob("*.fa");
+}
+
+foreach my $input_file (@files) {
+	open(in_file, $input_file)  || die "Error: File $input_file can not be opened";
+	my $output_file = $input_file . "_fixed.fa";
+	open(out_file, ">$output_file");
+	
+	my $species = find_species($input_file);
+
+	while(<in_file>) {
+		my $line = $_;
+		chomp $line;
+		if ($line =~ /^>/) {
+			my $gene = find_gene($line,$species);
+			$line = ">" . "$gene";
+		}
+		print out_file "$line\n";
+	}
+
+	close(in_file);
+	close(out_file);
+}
-- 
2.34.1