Alignements Smith-Waterman distribués avec SOAP::Lite et BioPerl

1. Problèmatiques

# XScreenSaver Preferences File
# Written by xscreensaver-demo 3.28 for olly on Thu Apr  4 13:35:55 2002.
# http://www.jwz.org/xscreensaver/

timeout:        0:10:00
cycle:          0:10:00
lock:           False
lockTimeout:    0:00:00
passwdTimeout:  0:00:30
visualID:       default
installColormap:True
verbose:        True
timestamp:      False
splash:         True
splashDuration: 0:00:05
demoCommand:    xscreensaver-demo
prefsCommand:   xscreensaver-demo -prefs
helpURL:        http://www.jwz.org/xscreensaver/man.html
loadURL:        netscape -remote 'openURL(%s)' || netscape '%s'
nice:           10
fade:           True
unfade:         False
fadeSeconds:    0:00:03
fadeTicks:      20
captureStderr:  True
font:           *-medium-r-*-140-*-m-*
dpmsEnabled:    False
dpmsStandby:    0:00:10
dpmsSuspend:    0:00:10
dpmsOff:        0:00:10
programs:       /home/olly/swc/swc.pl
pointerPollTime:0:00:05
windowCreationTimeout:0:00:30
initialDelay:   0:00:00
sgiSaverExtension:True
mitSaverExtension:False
xidleExtension: True
procInterrupts: True
overlayStderr:  True

2. Pré-requis

3. Créer la partie serveur

3.1 Base de données

mysql> create database SWC
    -> ;
Query OK, 1 row affected (0.56 sec)

drop table if exists SEQUENCES;
CREATE TABLE SEQUENCES (
ID       int not null auto_increment primary key,
ACCNUM   varchar(10),
SEQUENCE text
);

drop table if exists JOBS;
CREATE TABLE JOBS (
ID	int not null auto_increment primary key,
SEQID	int,
SEQST   int,
SEQEN   int
);

drop table if exists RESULTS;
CREATE TABLE RESULTS (
ACCNUM1	varchar(10),
ACCNUM2 varchar(10),
ALNSEQ1 text,
ALNSEQ2 text,
SCORE   float
);

#!/usr/bin/perl -w

use SOAP::Transport::HTTP;
use DBI();


 SOAP::Transport::HTTP::Daemon
    -> new (LocalAddr => 'localhost', LocalPort => 7777)
    -> dispatch_to('Server')
    -> handle;

package Server;

sub connect {
    
    $dbh = DBI->connect(    "DBI:mysql:database=SWC;host=localhost", 
                        "", 
                        "",
                        {'RaiseError' => 1}
			    );
    return $dbh;
}

sub getSequences {
    my ($class) = @_;
    
    my $CHUNK = 40; 

    $dbh = $class->connect();
    
    # get the total number of sequences in the DB
    $query = "SELECT count(*) AS COUNT FROM SEQUENCES";
    $sth = $dbh->prepare($query);
    $sth->execute;
    $count = $sth->fetchrow_hashref->{COUNT};
    
    print "got $count sequences in DB ...\n";

    
    # get the last job
    $query = "SELECT * FROM JOBS ORDER BY ID DESC LIMIT 1";
    $sth = $dbh->prepare($query);
    $sth->execute or die ("pb ..\n");
    $row = $sth->fetchrow_hashref;
    print "Issuing query $query\n";

    # if the END of the last job is equal to the number of seqs
    if (!$row) {
	$seqid = 1;
	$seqst = 1;
	$seqen = 1 + $CHUNK;
    } elsif ($row->{SEQEN} >= $count) {
	$seqid = $row->{SEQID} + 1;
	$seqst = 1;
	$seqen = 1 + $CHUNK;
    } else {
	$seqid = $row->{SEQID};
	$seqst = $row->{SEQEN} + 1;
	$seqen = ($row->{SEQEN} + $CHUNK > $count ? $count : $row->{SEQEN} + $CHUNK);
    }
    
    
    # create a new job
    $query = "INSERT INTO JOBS VALUES ('null', '$seqid', '$seqst', '$seqen')";

    print "Issuing query $query\n";
    $dbh->do($query) or die("pb ...\n");

    # get the target sequence
    $query = "SELECT * FROM SEQUENCES WHERE ID = $seqid";
    print "Issuing query $query\n";

    $sth = $dbh->prepare($query);
    $sth->execute;
    @seqs = ();
    $row = $sth->fetchrow_hashref;
    push(@seqs, $row);
    
    
    # get the sequences 
    $query = "SELECT * FROM SEQUENCES WHERE ID >= $seqst AND ID != $seqid LIMIT $CHUNK";
    print "Issuing query $query\n";

    $sth = $dbh->prepare($query);
    $sth->execute;
    
    while ($row = $sth->fetchrow_hashref) {
	push(@seqs, $row);
    }
    
    return \@seqs;
}

sub addResults {
    my ($class, $c) = @_;

    $dbh = $class->connect();

    print "results :\n";
    foreach $aln (@$c) {
	print $aln->{ACCNUM1} . "-" . $aln->{ACCNUM2} . "\n";
	$query = "INSERT INTO RESULTS VALUES ('" . $aln->{ACCNUM1} . "', '" .
	        $aln->{ACCNUM2} . "', '" .
		$aln->{ALNSEQ1} . "', '" .
	        $aln->{ALNSEQ2} . "', '" .
		$aln->{SCORE} . "')";
	#print "$query\n";
	$dbh->do($query) or die "oops ...\n";
    }
}

#!/usr/bin/perl



use Bio::Tools::pSW;
use Bio::Seq;
use Bio::SeqIO;
use Bio::AlignIO;
use GDBM_File;
use SOAP::Lite;


tie(my(%indexdb),   'GDBM_File',"sequences.db", &GDBM_WRCREAT, 0644);

tie(my(%varsdb),    'GDBM_File',"vars.db",      &GDBM_WRCREAT, 0644);

tie(my(%resultsdb), 'GDBM_File',"results.db",   &GDBM_WRCREAT, 0644);


sub handler{
   local ($signal) = @_;

   print "Got the $signal signal, aborting!\n";

   $varsdb{"theid"} = $theid;
   $varsdb{"j"} = $j;
   
   exit();
}

$SIG{'TERM'} = 'handler';


RETURN:

# list the stored results 
while (($k, $v) = each(%resultsdb)) {
    print ">$k\n";
    print "$v\n";
}


# in case the whole chunk has been computed

print scalar(keys(%resultsdb));
print scalar(keys(%indexdb));


if (scalar(keys(%resultsdb)) == scalar(keys(%indexdb)) - 1) {
    # send back the results
    @res = ();
    while (($k, $v) = each(%resultsdb)) {
	($a1, $a2) = $k =~ /^(.+)\|(.+)$/;
	($score, $s1, $s2) = $v =~ /^(.+)\|(.+)\|(.+)$/;
	my %hash = ("ACCNUM1" => $a1,
		  "ACCNUM2" => $a2,
		  "ALNSEQ1" => $s1,
		  "ALNSEQ2" => $s2,
		  "SCORE"   => $score );
	push(@res, \%hash);

    }
            
    $soap_response = SOAP::Lite
	 -> uri('http://localhost/Server')
	 -> proxy('http://localhost:7777')
	 -> addResults(\@res);

    %indexdb = ();
    %resultsdb = ();
    %varsdb = ();
    
}


if (scalar(keys(%indexdb)) == 0) {

    # get new sequences

    $soap_response = SOAP::Lite
	-> uri('http://localhost/Server')
        -> proxy('http://localhost:7777')
	-> getSequences();
    
    print "Getting data from the soap server !\n";
    $res = $soap_response->result;
    $cnt = 0;
    foreach $seq (@$res) {
	$theid = $seq->{ACCNUM} if ($cnt++ == 0);

	print $seq->{ACCNUM} . "\n";
	$indexdb{$seq->{ACCNUM}} = $seq->{SEQUENCE};
    }
    $thej = 0;
} else {
    
    # the sequence set is not empty, this could have been a SIGTERM
    $thej  = ($varsdb{"j"}?$varsdb{"j"}:0);
    $theid = $varsdb{"theid"};
}




@keys1 = keys %indexdb;


$factory = new Bio::Tools::pSW( '-matrix' => 'blosum62.bla',
	                           '-gap' => 12,
				   '-ext' => 2, );
											            

 LABEL: for ($j=$thej; $j<=$#keys1; $j++) {
     
     next LABEL if ($theid eq $keys1[$j]);
     print "Aligning " . $theid . " and " .  $keys1[$j] . " ($j)\n";
     $seq1    = Bio::Seq->new( -seq => $indexdb{$theid}, -id => $theid);
     $seq2    = Bio::Seq->new( -seq => $indexdb{$keys1[$j]}, -id => $keys1[$j]);
     
     $aln = $factory->pairwise_alignment($seq1, $seq2);
     
     @seqs = $aln->each_seq;
     
     $record = "";
     $cle    = "";
     
     $record .= $aln->percentage_identity;
     $record .= "|";
     $record .= $seqs[0]->seq;
     $record .= "|";
     $record .= $seqs[1]->seq;
     
     $cle    .= $seqs[0]->id;
     $cle    .= "|";
     $cle    .= $seqs[1]->id;
     
     $resultsdb{$cle} = $record;

 }


goto RETURN;

untie %indexdb;
untie %varsdb;
untie %resultsdb;