# prom5.pl (GetPromoters)
#
# Daniel Westreich, July 2002
# Mullins Lab, Microbiology Dept.
# University of Washington
#
# Usage:	prom5.pl filename
#

use IO::Socket;

$RANGE_BACK = 4300;
$RANGE_FRONT = 199;

$host = "www.ncbi.nlm.nih.gov";

$promfirst="http://www.ncbi.nlm.nih.gov/entrez/viewer.fcgi?cmd=&txt=on&save=&cfm=&_from=";
$prommid1="&_to=";
$prommid2="&_strand=";
$prommid3="&list_uids=";
$promend="&db=&view=fasta&__from=&__to=&__strand=";

open(LIST, ">promoterlist.out");

while(<>) {
    if (/^\#/) #comments
    {
	next;
    }
    chomp($_);
    @words= split(/\|/);
    if (@words == 0)
    {
	next;
    }
    $gene = @words[0];
    $gi = @words[1];
    $strand = @words[2];
    $address = @words[3];
    
    print "For gene ".$gene."\n";

    GetPromoterSequences($gi,$strand,$address,$gene);
    print LIST $gene."_promoter.out\n";
}

######################

sub GetPromoterSequences {
    $gi = $_[0];
    $strand = $_[1];
    $address = $_[2];
    $gene = $_[3];
    
    $outfile = ">".$gene."_promoter.out";
    open(OUT, $outfile);

    if ($strand eq "Crick")
    {
	$strand = 2;
	$begaddress = (int($address) - $RANGE_FRONT);
	$endaddress = (int($address) + $RANGE_BACK);
    }
    else
    {
	$strand = 0;
	$begaddress = (int($address) - $RANGE_BACK);
	$endaddress = (int($address) + $RANGE_FRONT);
    }

    $docurl = $promfirst.$begaddress.$prommid1.$endaddress.$prommid2.$strand.$prommid3.$gi.$promend;
#    print ($docurl);
    $cSeqs = 0;

    $remote = IO::Socket::INET->new( Proto => "tcp", PeerAddr => $host, PeerPort => "http(80)", );
    unless($remote)  { die "cannot connect to http daemon on $host" }
    $remote->autoflush(1);
    print $remote "GET $docurl HTTP/1.0\n\n";
    
#   DEBUG
#   print OUT $docurl."\n\n"; 
    while (<$remote>)
    {
	if (/\>gi\|/)
	{
	    print OUT;
	    last;
	}
    }
    while (<$remote>)
    {
	print OUT;
    }
    -close $remote;  
}