Rename BACs

From RAJ INFO
Jump to navigation Jump to search
#! /usr/bin/perl -w
 
# *************************************************************
# This Programme does the following in sequence
# -> Reads the file in a folder
# -> Reads the fasta, name appropriately and writes a concatenated outfile
# Author : Rajkumar (itc@rajkumar.in)
# Release: APR 2010
# *************************************************************
 
# Libraries
	use strict;
	use Bio::Perl;
	use Bio::SearchIO;
	use Parallel::ForkManager;

 
# At the top: Time Start
	my ($sec,$min,$hour,$mday,$mon,$year,$wday,$yday,$isdst) = localtime(time);
 
# Variables
	my $path    = "/home/raj/bio";
	my $bacdir  = "$path/tobacco/data/BAC";
	my $outfile = "$path/tobacco/data/bac.fa";
	
	my $pm = new Parallel::ForkManager(1);
	 
	my ($i)    = $_;
	my (@files)= @_;
 


# Clear the Files
	
	if (-e "$outfile") {`rm $outfile`;}

 
# Function
 
	opendir (DIR, "$bacdir") or die $!;
	@files = readdir(DIR);
	close(DIR);
 
	foreach my $file(@files) {
		next if ($file eq "."  or $file eq "..");
		$pm->start and next;


		my $in  = Bio::SeqIO->new(-file => "$bacdir/$file", -format => 'Fasta');
		my $identifier;
		my $sequence;
 
		while ( my $seq = $in->next_seq() ) {
			$identifier = $seq->id;
	  		$sequence = $seq->seq;
	  		
	  		$file =~ s/-assembly.bases//;
	  		$identifier = "BAC_".$file."_".$identifier;

			my ($fasta) = $_;
			$fasta = ">$identifier\n$sequence\n";
			
			open (F, ">>".$outfile);
			print F $fasta;
			close F;

		}
		$pm->finish;
		
	}

	$pm->wait_all_children;

	print "\n\tAll Done!\n\n";
	
 
# At the End: Time Calculation
	my ($sec_,$min_,$hour_,$mday_,$mon_,$year_,$wday_,$yday_,$isdst_) = localtime(time);
	my $days  = $mday_-$mday;
	my $hours = $hour_-$hour;
	my $mins  = $min_-$min;
	my $secs  = $sec_-$sec;
	print "\n------------------------\nDuration: ";
	if ($days>0) { print "$days days";}
	if ($hours>0) { print ": $hours hours";}
	if ($mins>0) { print ": $mins mins";}
	if ($secs>0) { print ": $secs secs";}
	print "\n------------------------\n";
 
# END