From 6a7a9cb48f41a72843d110f7fd58fa9ce5da5dc0 Mon Sep 17 00:00:00 2001
From: Hirokazu Chiba <hchiba1@gmail.com>
Date: Thu, 17 Oct 2024 19:14:32 +0900
Subject: [PATCH] Add bin/extract_taxid.pl

---
 bin/extract_taxid.pl | 40 ++++++++++++++++++++++++++++++++++++++++
 1 file changed, 40 insertions(+)
 create mode 100755 bin/extract_taxid.pl

diff --git a/bin/extract_taxid.pl b/bin/extract_taxid.pl
new file mode 100755
index 0000000..781ff6d
--- /dev/null
+++ b/bin/extract_taxid.pl
@@ -0,0 +1,40 @@
+#!/usr/bin/perl -w
+use strict;
+use File::Basename;
+use Getopt::Std;
+my $PROGRAM = basename $0;
+my $USAGE=
+"Usage: $PROGRAM
+";
+
+my %OPT;
+getopts('', \%OPT);
+
+my %HASH;
+while (<>) {
+    chomp;
+    if (/^#/) {
+        next;
+    }
+
+    my @f = split(/\t/, $_, -1);
+    if (@f != 5) {
+        die;
+    }
+
+    if ($f[0] !~ /^\d+$/) {
+        die;
+    }
+    my $taxid1 = $f[0];
+    $HASH{$taxid1} = 1;
+
+    if ($f[2] ne "Ortholog") {
+        die;
+    }
+    my $taxid2 = $f[3];
+    $HASH{$taxid2} = 1;
+}
+
+for my $taxid (sort {$a <=> $b} keys %HASH) {
+    print "$taxid\n";
+}