webblast.pl 33.7 KB
Newer Older
Sebastien Moretti's avatar
Sebastien Moretti committed
1
2
3
#!/usr/bin/env perl
#
#
4
#date : 2007/11/19
Sebastien Moretti's avatar
Sebastien Moretti committed
5
6
7
8
9
10
11
12
13
14
15
16
17
18
#prog : webblast.pl
#subj : make a BLAST/WU-BLAST (by HTTP request or locally) against a database with a file containing sequences in fasta format
####### method genid, pdbid and profile 
#
#############################################################################################
#use Env qw(HOME);
#use lib "$HOME/.lib_webblast/";
use LWP::UserAgent; 
use HTML::Parser;                                            # @@@@@@@ #       
use HTTP::Request::Common qw(POST);                         # @/^   ^\@ # 
use URI::Escape;                                           # @/ -   - \@ #
use Getopt::Long;                                         ##  \   ^   /  ##
use strict;                                              ##    |  0  |    ##
use warnings;                                           ####### \ _ / #######
19
##############################################################################################
Sebastien Moretti's avatar
Sebastien Moretti committed
20
21


22
23
24
25
26
27
28
29
############################  EXPRESSO PARAM  ##########################################################
my $database_expresso  = 'pdb';                                            #PDB database name          #
my $blast_dir_expresso = '/mnt/local/bin/blastall';                        #blastall executable        #
my $BLASTMAT           = 'export BLASTMAT=/mnt/local/ncbi/data/';          #matrix directory for blast #
my $BLASTDB            = 'export BLASTDB=/scratch/frt/blastnet/database/'; #PDB directoty              #
########################################################################################################
    
my $runblast = '/mnt/local/bin/runblast.pl';
Sebastien Moretti's avatar
Sebastien Moretti committed
30

31
my(@list_encoded)=(), my(@list_pdb)=(), my(%deja_vu)=(), my(@pdb_list)=(),my($i)=0, my(@names)=(), my $locale=0, my $distant=0, my $database, my $blast_way;
Sebastien Moretti's avatar
Sebastien Moretti committed
32
33
my($ua)= LWP::UserAgent->new;

34

Sebastien Moretti's avatar
Sebastien Moretti committed
35
36
37
38
39
40
41
42
43
##-- Variables d'environnements 

my($database_var)= $ENV { 'DATABASE' };
my($blast_var)   = $ENV { 'BLAST_DIRECTORY' };

##-- Recupere Options/parametres du BLAST && controle des options ds OPTIONS_GET
 
my($program,$database_line,$blast_line,$query_file,
   $out_file,$identity_treshold,$cover_tresh,$Eval,
44
   $align,$matrix,$filter,$method,$orgn,$process,$quiet,$gigablast)= &OPTIONS_GET();
Sebastien Moretti's avatar
Sebastien Moretti committed
45

46
##-- Determination BLAST LOCAL /DISTANT && Controle database/programme
Sebastien Moretti's avatar
Sebastien Moretti committed
47

48
49
unless (-e $query_file ) { print {*STDERR} "\nfile does not exist!\n";exit;}
unless (-s $query_file ) { print {*STDERR} "\nyour file is empty!\n";exit; }
Sebastien Moretti's avatar
Sebastien Moretti committed
50

51
if ((($database_line || $database_var) && ($blast_line || $blast_var)) ) 
Sebastien Moretti's avatar
Sebastien Moretti committed
52
{   
53
54
55
56
57
58
59
60
61
62
63
64
65
66
    if ($database_line=~/expressopdb/ && $blast_line=~/blastexpresso/)
    {
	#mode special pour fichier de configuration du serveur Expresso
	$locale=1;
	$database="expressopdb";
	unless ($quiet=~ /on/i) { print {*STDERR} "\nRUN BLAST LOCALY\n"; }
    }
    else
    {
	($database)=$database_line || $database_var;	
	my($blast_tp)=$blast_var || $blast_line; $locale=1;
	$blast_way = &CONTROLE_DB_PG($database,$blast_tp,$program);
	unless ($quiet=~ /on/i) { print {*STDERR} "\nRUN BLAST LOCALY\n"; }
    }
Sebastien Moretti's avatar
Sebastien Moretti committed
67
}
68

Sebastien Moretti's avatar
Sebastien Moretti committed
69
70
else
{
71
72
73
    $database = &NCBI_DATABASE($database_line); $distant=1;    
    if ($gigablast=~ /^yes$/i) { $locale=2; $distant=0; unless ($quiet=~ /on/i) { print {*STDERR} "\nRUN GIGABLASTER\n"; }}   
    else { unless ($quiet=~ /on/i)  {print {*STDERR} "\nRUN BLAST AT THE NCBI\n";}} 
Sebastien Moretti's avatar
Sebastien Moretti committed
74
75
}
 
76
##-fixation de parametres selon la valeur du flag -method
Sebastien Moretti's avatar
Sebastien Moretti committed
77

78
if ($method=~ /^pdbid$/i)
Sebastien Moretti's avatar
Sebastien Moretti committed
79
{   
80
81
82
83
84
85
86
  
    if    ($gigablast=~ /^yes$/i)          
    {
	if ($database ne "pdb") { print {*STDERR} "\nprovide a valid database name FOR RUN GIGABLASTER: nr,pdb or refseq_protein\n";exit;} else {$database="pdbaa";}
    }
    elsif ($gigablast=~ /^no$/i && $distant==1) { $database="pdb";}
          
Sebastien Moretti's avatar
Sebastien Moretti committed
87
}
88
89
90
elsif ($method=~ /^geneid$/i)
{
    if ($distant==1) { unless ($database eq "nr" || $database eq "swissprot" || $database eq "pdb") { $database="refseq_protein";}}   
Sebastien Moretti's avatar
Sebastien Moretti committed
91
}
92
93
94
95
96
97
98
99
elsif ($method=~/^profile$/i)
{
    if ($distant==1) {	unless ($database eq "nr" || $database eq "swissprot" || $database eq "refseq_protein" )  { $database="pdb"; } }
}
else { die "unknown method\n";}

if (($orgn !~ /All\+organisms/) && ($locale=~/1|2/))
   { print {*STDERR} "-organism option can't be used locally or with -gigablast option!\n";exit;}
Sebastien Moretti's avatar
Sebastien Moretti committed
100

Sebastien Moretti's avatar
Sebastien Moretti committed
101
102
103
104
105
##---PRINT option values
unless ( $quiet =~ /on/i ){
    print {*STDERR} "

             Program :  $program
Sebastien Moretti's avatar
Sebastien Moretti committed
106
             Database : $database
Sebastien Moretti's avatar
Sebastien Moretti committed
107
108
             Method :   $method

Sebastien Moretti's avatar
Sebastien Moretti committed
109
             Query_file : $query_file
Sebastien Moretti's avatar
Sebastien Moretti committed
110
111
112
113
114
115
             Out_file :   $out_file
";
    print {*STDOUT} "
             Evalue threshold :         $Eval
             Matrix :                   $matrix
             Filter :                   $filter
Sebastien Moretti's avatar
Sebastien Moretti committed
116
             Blast_identity_threshold : $identity_treshold
Sebastien Moretti's avatar
Sebastien Moretti committed
117
118
             Cover threshold :          $cover_tresh
";
119
    print {*STDERR} "
Sebastien Moretti's avatar
Sebastien Moretti committed
120
121
122
123
124
125
126
127
128
129
130
131
             Number of hits :            $align
             Number of processors used : $process
";
    if ( $gigablast=~ /^yes$/i ){
        print {*STDERR} "\n             gigablast: yes\n";
    }
    unless ($locale) {
        print {*STDERR} "\n             Organism : $orgn\n";
    }


    print {*STDERR} "\n***************************************************************\n\n";
Sebastien Moretti's avatar
Sebastien Moretti committed
132
133
}

134

Sebastien Moretti's avatar
Sebastien Moretti committed
135
136
#-- LOCAL/DISTANT BLASTP
   
137
if    ($locale==1 || $locale==2) {@list_pdb= &LOCAL_BLAST ($blast_way,$database,$query_file,$Eval,$align,$method,$matrix,$filter,$process,$gigablast,$database_expresso,$blast_dir_expresso,$runblast);   }
Sebastien Moretti's avatar
Sebastien Moretti committed
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
elsif ($distant==1)  { @list_pdb= &WEB_BLAST   ($query_file,$Eval,$program,$database,$matrix,$method,$align,$orgn,$filter); }
else              { die " Report bug to armougom\@igs.cnrs-mrs.fr\n"}; 

#-- PARSE BLAST RESULTS -> MAKE A PDB_ID LIST
if ($method =~ /^pdbid$/i) 
{ 
    my(@result_sort)= &PARSING (\@list_pdb,$locale,$distant,$method,$quiet,$database,$gigablast);
   
                      &AFFICHAGE_PDB_PARSING (\@result_sort,$cover_tresh,$identity_treshold,$out_file);
                      exit;
}

#-- PARSE BLAST RESULT -> MAKE LIST OF REFSEQ ID
elsif ($method =~ /^geneid$/i)
{ 
    my(@result_sort)= &PARSING (\@list_pdb,$locale,$distant,$method,$quiet,$database,$gigablast);
                      &AFFICHAGE_REFSEQ_PARSING (\@result_sort,$cover_tresh,$identity_treshold,$out_file);
                      exit;
}

#-- PARSE BLAST RESULT -> MAKE PROFILE
159
elsif ($method=~ /^profile$/i) { &PROFILE (\@list_pdb,$out_file,$distant); exit; }
Sebastien Moretti's avatar
Sebastien Moretti committed
160
else { die " \nFATAL ERROR :  Method or database error\n" ;}
161

Sebastien Moretti's avatar
Sebastien Moretti committed
162
exit;
163
                   
Sebastien Moretti's avatar
Sebastien Moretti committed
164
165
166
                                               ##############
###############################################  FONCTIONS  ####################################################################
                                              ##############
Sebastien Moretti's avatar
Sebastien Moretti committed
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
sub CONTROLE_DB_PG {
    my ($database, $blast_dir, $program) = @_;

    if ( !-e "$database" && !-e "$database.pin"){
        die "$database file does not exist\n";
    }
    if ( -d $database ){
        die "$database must be a file, not a directory\n";
    }
    if ( $blast_dir !~ /\/$/ ){
        $blast_dir .= '/';
    }
    my ($blastall) = $blast_dir.'blastall';
    if ( !-e $blastall){
        die "$blastall program not found \n";
    }
Sebastien Moretti's avatar
Sebastien Moretti committed
183

184
    return ($blastall);
Sebastien Moretti's avatar
Sebastien Moretti committed
185
186
187
}

#-------------------------------------------------------------------------------------------------------------------------------------
Sebastien Moretti's avatar
Sebastien Moretti committed
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
sub NCBI_DATABASE {
    my ($ncbi_db) = @_;

    my (%all_db) =
    (
            'nr'             =>'1',
            'pdb'            =>'1',
            'swissprot'      =>'1',
            'refseq_protein' =>'1',
    );

    if ( exists $all_db{$ncbi_db} ){
        return($ncbi_db);
    }
    elsif ( $ncbi_db eq '' ){
        return ('');
    }
    else{
        return (1);
    }
Sebastien Moretti's avatar
Sebastien Moretti committed
208
209
210
211
} 
#------------------------------------------------------------------------------------------------------------------------
sub HELP
{   
Sebastien Moretti's avatar
Sebastien Moretti committed
212
213
    my ($org, @orga) = &LIST_ORGA();
    my ($list_orga)  = join(', ', @orga);
Sebastien Moretti's avatar
Sebastien Moretti committed
214

215
216
    print {*STDERR} "
                      usage: $0 -infile <fasta file> -method <pdbid/geneid or profile> options []\n
Sebastien Moretti's avatar
Sebastien Moretti committed
217
218

                            -program ...... Program Name (blastp)
Sebastien Moretti's avatar
Sebastien Moretti committed
219
                                            Default = blastp
220
                            -database ..... Database at NCBI (nr, pdb, swissprot, refseq_protein) or indicate a local fasta file
Sebastien Moretti's avatar
Sebastien Moretti committed
221
222
223
224
                                            Default = pdb at NCBI
                            -infile ....... Query_file = a list of sequences in fasta format
                            -outfile ...... Name the outfile to make a template file for t_coffee
                                            Default = STDOUT or default.profile if method is profile
225
226
                            -evalue ....... Evalue threshold Default = 1
                            -matrix ....... PAM30 PAM70 BLOSUM45 BLOSUM80
Sebastien Moretti's avatar
Sebastien Moretti committed
227
                                            Default BLOSUM62
228
229
                            -method ....... geneid, pdbid, profile
                            -gigablast..... yes/no FASTER REMOTE BLAST with Gigablaster
Sebastien Moretti's avatar
Sebastien Moretti committed
230
                                            (Stephane Audic program: http://www.igs.cnrs-mrs.fr/Giga2/~database/remoteblast.cgi)
Sebastien Moretti's avatar
Sebastien Moretti committed
231
                                            Default no
232
                            -filter ....... T or F locally, L or R or M or C or V for distant blast
Sebastien Moretti's avatar
Sebastien Moretti committed
233
                                            Default = Off
234
235
                            -organism ..... $list_orga are available
                                            Default is All_organisms
Sebastien Moretti's avatar
Sebastien Moretti committed
236
                            -identity ..... blast identity threshold = provide a % for view only the results upper or equal to the threshold
237
238
239
                                            Default 50
                            -cover ........ Cover threshold = provide a % : sequence covering Default: 30
                            -hits ........  Number of hits
Sebastien Moretti's avatar
Sebastien Moretti committed
240
                                            Default = 1
241
242
243
244
245
                            -processor .... Number of processors to use
                                            Default = 1
                            -blast_dir .... Indicates where your BLAST directory is installed localy
                            -quiet ........ on : do not display all the default/defined blast parameters
                                            Default off
Sebastien Moretti's avatar
Sebastien Moretti committed
246
247
248
249
250

                     Environement Variables
                     These variables can be set from the environement
           DATABASE......................[Indicates where your database file must be fetched (localy)]
           BLAST_DIRECTORY...............[Indicates where your BLAST directory is installed localy]
251

Sebastien Moretti's avatar
Sebastien Moretti committed
252
253
254
255
256
257
";
	
    exit;
    
}
#-----------------------------------------------------------------------------------------
Sebastien Moretti's avatar
Sebastien Moretti committed
258
259
sub OPTIONS_GET {   
    my %opt = ();
Sebastien Moretti's avatar
Sebastien Moretti committed
260
261
262
 
    GetOptions 
              (
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
	       'infile=s'    =>\$opt{infile},
	       'outfile=s'    =>\$opt{outfile},
	       'program=s'     =>\$opt{program},
	       'database=s'     =>\$opt{database},
	       'blast_dir=s'     =>\$opt{blast_dir},
	       'identity=f'       =>\$opt{treshold},
	       'cover=f'           =>\$opt{cover},
	       'evalue=f'           =>\$opt{evalue},     
	       'hits=i'              =>\$opt{hits},	   
	       'matrix=s'             =>\$opt{matrix},
	       'filter=s'              =>\$opt{filter},
	       'method=s'               =>\$opt{method},
           'organism=s'              =>\$opt{organism},
	       'processor=i'              =>\$opt{processor},
	       'quiet=s'                   =>\$opt{quiet},
	       'gigablast=s'                =>\$opt{gigablast},
Sebastien Moretti's avatar
Sebastien Moretti committed
279
280
281
282
283
284
	       );
  
    if ($ARGV[0]) {print "Unprocessed by Getopt::Long\n $ARGV[0]\n"; &HELP();} 
   
  

285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
    my($evalue_tresh) = $opt{'evalue'};       unless ($evalue_tresh) { $evalue_tresh=1;};
    my($cover_tresh)  = $opt{'cover'};         unless (defined $cover_tresh)  { $cover_tresh=30;};
    my($query_file)   = $opt{'infile'};         unless ($query_file)   { print {*STDERR} "Flag -infile must be defined\n"; &HELP();};
    my($outfil)       = $opt{'outfile'};         unless ($outfil)       { $outfil='';};
    my($treshold)     = $opt{'treshold'};         unless (defined $treshold)     { $treshold=50;};
    my($blast_dir)    = $opt{'blast_dir'};         unless ($blast_dir)    { $blast_dir='';};
    my($database)     = $opt{'database'};           unless ($database)     { $database='';};
    my($program)      = $opt{'program'};             unless ($program)      { $program='blastp';};  
    my($align)        = $opt{'hits'};                 unless (defined $align)       { $align=1;};  
    my($matrix)       = $opt{'matrix'};                unless ($matrix)      { $matrix='BLOSUM62';};
    my($filter)       = $opt{'filter'};                 unless ($filter)      { $filter='F';};
    my($method)       = $opt{'method'};                  unless ($method)      {print {*STDERR} "Flag -method must be defined\n"; &HELP();};
    my($organism)     = $opt{'organism'};                 unless ($organism)    { $organism='All organisms';};
    my($process)      = $opt{'processor'};                 unless ($process)     { $process=1;};   
    my($param)        = $opt{'quiet'};                      unless ($param)       { $param='off';}; 
    my($gigablast)    = $opt{'gigablast'};                   unless ($gigablast)    {$gigablast='no';}; 
    if ($method !~ /(^geneid$|^pdbid$|^profile$)/i)  { print {*STDERR} "unknown method for the flag -method\n";&HELP(); }

	if ($treshold <0 || $treshold >100)               { print {*STDERR} "\nout of range for the option -treshold \n"; &HELP();}  
	if ($cover_tresh <0 || $cover_tresh >100)         { print {*STDERR} "\nout of range for the option -cover \n"; &HELP();} 
	if ($align <0)                                    { print {*STDERR} "\n error with option   align\n"; &HELP();}
	if ($gigablast!~/^yes$|^no$/i)                    { print {*STDERR} "invalid argument for gigaglast option : yes/no\n";exit;};
	if ($filter!~ /^[TFRLMCV]{1}$|^off$/i)                  {print {*STDERR}  "valid values for -filter are T,F,R,L,M,C,or V!\n";exit;}
	if ($matrix!~ /PAM30|PAM70|BLOSUM45|BLOSUM80|BLOSUM62/) { print {*STDERR} "valid values for -matrix  are PAM30,PAM70,BLOSUM45,BLOSUM80 or BLOSUM62\n";exit }
	if ($outfil eq "" && $method=~ /^profile$/i) { $outfil='default_profile.template'}
	
	if ($param!~ /^on$|^off$/i)                       { print {*STDERR} "valid values for -quiet is on or off\n";exit;}
	my($orgn,@all_orgn)= &ORGN($organism);
	return ($program,$database,$blast_dir,$query_file,
		    $outfil,$treshold,$cover_tresh,$evalue_tresh,
		    $align,$matrix,$filter,$method,$orgn,$process,$param,$gigablast);
Sebastien Moretti's avatar
Sebastien Moretti committed
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
}

#--------------------------------------------------------------------------------------------------
sub RECOVER
{
    my($pdb_result,$aln_length,$length_query)=@_;
    my $nb_gap=0;
    
    if ($pdb_result=~ /(score.+?\n\n\n).+?score/ism) #cas ou plusieurs HSP, prend que le 1er
    { $pdb_result= $1;}
    
    $length_query  =~ s/,//g;
    my ($requete)  =  join('',($pdb_result=~/^Query(.*)\n/gm));
    $requete       =~ s/[^A-Z-]//g;
    my(@sequence)  =  split('',$requete);
 
    for (my $i=0; $i<=$#sequence; $i++)
    {
	if($sequence[$i] eq "-"){ ++$nb_gap; }
    }
    my($recouvrement)= sprintf("%-3d",(($aln_length-$nb_gap)/$length_query)*100);
    undef(@sequence);
    return ($recouvrement,$nb_gap); 
   
}
#--------------------------------------------------------------------------------------------------
sub WEB_BLAST
{
344
    open (SOR1, '>', 'web_tempo.result') or die;
Sebastien Moretti's avatar
Sebastien Moretti committed
345
    my($query_file,$Eval,$program,$database,$matrix,$method,$align,$orgn,$filter)=@_;
346
    my $aln_view, my $format='Txt';
Sebastien Moretti's avatar
Sebastien Moretti committed
347
    my($description)=$align;
348
    if ($method=~/^profile$/i) { $aln_view ='FlatQueryAnchoredNoIdentities'} else { $aln_view ='Pairwise'}
Sebastien Moretti's avatar
Sebastien Moretti committed
349
   
350
    if ($filter eq 'F') { $filter='off';}
Sebastien Moretti's avatar
Sebastien Moretti committed
351

352
    $/='>';
Sebastien Moretti's avatar
Sebastien Moretti committed
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
    open(FIC,$query_file) or die "can not open $query_file $!\n";
    my(@sequences)=<FIC>;
    close FIC;
    shift(@sequences);

    foreach my $sequence(@sequences)	
    {
	$sequence=~ s/>//g;
	$sequence=">$sequence";
	
	my($name)=($sequence=~ /^>(.+)\n/);
	push(@names, $name); 
	my($encoded_query)= uri_escape($sequence);
	push (@list_encoded, $encoded_query);    
    } 
    
    undef(@sequences);
    if (scalar (@names != @list_encoded)) { die "error $!";}     
    foreach my $encoded_seq(@list_encoded)
    {    
	my $nb=0;
374
	print {*STDERR} "BLAST $names[$i]...";
Sebastien Moretti's avatar
Sebastien Moretti committed
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
	
#-- BUILD THE REQUEST
		
	my($arguments) = "CMD=Put&ENTREZ_QUERY=$orgn&CDD_SEARCH=off&FILTER=$filter&MATRIX_NAME=$matrix&PROGRAM=$program&DATABASE=$database&QUERY=" . $encoded_seq;
	
	my($req) = new HTTP::Request POST => 'http://www.ncbi.nlm.nih.gov/blast/Blast.cgi';
	$req -> content_type('application/x-www-form-urlencoded');
	$req -> content($arguments);
	
#-- GET THE RESPONSE : PARSE OUT THE REQUEST ID and THE ESTIMATED TIME
	my($response) = $ua -> request($req);
	
	if ($response -> content =~ /Server Error/i) { die "Server Error at NCBI!!Sorry try later\n"; }
	$response -> content =~ /^\s{4}RID = (.*)$/m;   my($rid) = $1;
	$response -> content =~ /^\s{4}RTOE = (.*)$/m;	my($wait)= $1;
	unless ($rid && $wait)             { die "parse error: $!" };
391
	for (my $j=0; $j<=$wait/2; $j++)   {    print {*STDERR} ".";	sleep 2;   }
Sebastien Moretti's avatar
Sebastien Moretti committed
392
393
394
395
396
	
	my($verif)=0;
	
	while ()
	{  		
397
		for (my $j=0; $j<=5; $j++)  { print  {*STDERR} ".";  sleep 1; }
Sebastien Moretti's avatar
Sebastien Moretti committed
398
399
400
401
402
		
		$req = new HTTP::Request GET =>
		    "http://www.ncbi.nlm.nih.gov/blast/Blast.cgi?CMD=Get&FORMAT_OBJECT=SearchInfo&RID=$rid";   
		$response = $ua->request($req);	   
		if    ($response->content =~ /Status=WAITING/im) {  next; }	
403
404
		elsif ($response->content =~ /Status=FAILED/im)  { print {*STDERR} "Search $rid failed\n"; $verif=1; last; }	    
		elsif ($response->content =~ /Status=UNKNOWN/im) { print {*STDERR} "Search $rid expired\n"; $verif=1; last; }	    
Sebastien Moretti's avatar
Sebastien Moretti committed
405
406
407
		elsif ($response->content =~ /Status=READY/im) 
		{	       
		    if   ($response->content =~ /ThereAreHits=yes/im){last;}	       
408
		    else { print {*STDERR} "No hits found.\n";$verif=1;last;  }
Sebastien Moretti's avatar
Sebastien Moretti committed
409
410
411
		}
		elsif ($response->content =~ /can\'t connect/im)
		{ 
412
		    print {*STDERR} "\nCan't connect to www.ncbi.nlm.nih.gov:80...new attempt"; 
Sebastien Moretti's avatar
Sebastien Moretti committed
413
		    if ($nb <3) { ++$nb; next; } 
414
		    else { print {*STDERR} "sorry, BLAST $names[$i] failed after 3 attempts!!\n"; $verif=1; last;}
Sebastien Moretti's avatar
Sebastien Moretti committed
415
		}
416
		else { print {*STDERR} "unknown error\n"; $verif=1; last; }
Sebastien Moretti's avatar
Sebastien Moretti committed
417
418
419
420
421
422
423
424
425
426
427
428
	    } 
	
	if($verif==1){ ++$i; next; }
	
#-- GET RESULT
	
	while ()
	{
	    sleep 3;
	    $req = new HTTP::Request GET => "http://www.ncbi.nlm.nih.gov/blast/Blast.cgi?CMD=Get&FORMAT_TYPE=$format&FILTER=off&EXPECT=$Eval&ALIGNMENTS=$align&DESCRIPTIONS=$align&ALIGNMENT_VIEW=$aln_view&RID=$rid";
		$response = $ua -> request($req);
	    
429
	    if   ($response->content =~ /Altschul/i) {  print {*STDERR} "Search Complete\n"; push(@list_pdb,$response -> content);last; }
Sebastien Moretti's avatar
Sebastien Moretti committed
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
	    else { next; }
	}
	print SOR1 (@list_pdb);
	++$i;
    }
    
    undef (@list_encoded);
    
    close SOR1;
    return (@list_pdb);
    
}
#-----------------------------------------------------------------------------------------------------------------------
sub LOCAL_BLAST
{
Sebastien Moretti's avatar
Sebastien Moretti committed
445
446
447
448
449
450
    my ($blast_dir, $database, $query_file, $Eval, $align,
        $method, $matrix, $filter, $process, $gigablast,
        $database_expresso, $blast_dir_expresso, $runblast) = @_;
    my $n = 0;
    if ( $method=~ /^profile$/i && $gigablast=~ /^no$/i ){ 
        open (COM,"$blast_dir -p blastp -d $database -i $query_file -m 6 -M $matrix -v $align -b $align -F $filter -e $Eval -a $process|") or die;
Sebastien Moretti's avatar
Sebastien Moretti committed
451
    }
Sebastien Moretti's avatar
Sebastien Moretti committed
452
    elsif ( $method=~ /^geneid$/i && $gigablast=~ /^no$/i ) 
Sebastien Moretti's avatar
Sebastien Moretti committed
453
    { 
Sebastien Moretti's avatar
Sebastien Moretti committed
454
        open (COM,"$blast_dir -p blastp -d $database -i $query_file  -v $align  -b $align -F $filter -M $matrix -e $Eval -a $process|") or die;
Sebastien Moretti's avatar
Sebastien Moretti committed
455
    }
Sebastien Moretti's avatar
Sebastien Moretti committed
456
    elsif ( $method=~ /^geneid$|^pdbid$/i && $gigablast=~ /^yes$/i )
Sebastien Moretti's avatar
Sebastien Moretti committed
457
    {
Sebastien Moretti's avatar
Sebastien Moretti committed
458
459
460
461
462
463
464
        unless ( $database eq "nr" || $database eq "pdb" || $database eq "refseq_protein" || $database eq "pdbaa" ){
            print {*STDERR} "\nsorry invalid database for gigablast\n";exit 1;
        }
        if ( $database eq 'pdb')            { $database = 'pdbaa';}
        if ( $database eq 'refseq_protein') { $database = 'refprot';}
        if ( $database eq '')               { print {*STDERR} "provide a valid database!\n"; exit;}
        open (COM,"$runblast -d $database -p blastp  -e $Eval -v $align -F F \<$query_file |");
465
    }
Sebastien Moretti's avatar
Sebastien Moretti committed
466
467
468
469
470
    elsif ( $method=~ /^profile$/i && $gigablast=~ /^yes$/i){
        print {*STDERR} "\nSorry method profile can't be used with -gigablast option\n";exit;
    } 
    elsif ( $method=~ /^pdbid$/i && $database=~ /expressopdb/ ){	
    	#BLAST pour Expresso
Sebastien Moretti's avatar
Sebastien Moretti committed
471
        open (COM,"$BLASTMAT; $BLASTDB; $blast_dir_expresso -p blastp -d $database_expresso -i $query_file -F $filter -e $Eval -M $matrix -v $align -b $align |") or die;       
Sebastien Moretti's avatar
Sebastien Moretti committed
472
    }
Sebastien Moretti's avatar
Sebastien Moretti committed
473
474
    else{  
        open (COM,"$blast_dir -p blastp -d $database -i $query_file -v 1 -b 1 -F $filter -e $Eval -M $matrix -v $align -b $align -a $process |") or die;
Sebastien Moretti's avatar
Sebastien Moretti committed
475
    }
476
    
Sebastien Moretti's avatar
Sebastien Moretti committed
477
    unless ( $quiet=~ /on/ ) { print {*STDERR} "\nrun BLAST..."; } 
478
    
Sebastien Moretti's avatar
Sebastien Moretti committed
479
    my ($name_database, $posted, $version) = ('', '', '');
Sebastien Moretti's avatar
Sebastien Moretti committed
480

Sebastien Moretti's avatar
Sebastien Moretti committed
481
    open (my $SOR2, '>', 'blast_result.txt') or die;
Sebastien Moretti's avatar
Sebastien Moretti committed
482
    
Sebastien Moretti's avatar
Sebastien Moretti committed
483
484
485
486
487
488
    $/ = 'Query=';
    while (<COM>){
        if ( $_=~ /Database:\s+(\S+)/g )     { $name_database = $1;}
        if ( $_=~ /Posted date:\s+(.+?)\n/ ) { $posted        = $1;}
        if ( $_=~ /(BLASTP\s+\S+)/o )        { $version       = $1;}
        print {$SOR2} $_;
Sebastien Moretti's avatar
Sebastien Moretti committed
489
490
491
492
        push (@list_pdb, $_) ; 
        if ( $_=~ /\s*(.+?)\s/ ){
            print {*STDERR} "\n$1 done";
        } 
Sebastien Moretti's avatar
Sebastien Moretti committed
493
494
    }
    close COM;
Sebastien Moretti's avatar
Sebastien Moretti committed
495
    close $SOR2;
496
    print {*STDERR} "\n";
Sebastien Moretti's avatar
Sebastien Moretti committed
497

Sebastien Moretti's avatar
Sebastien Moretti committed
498
    $name_database = $database if ( $name_database =~ m{/} );
Sebastien Moretti's avatar
Sebastien Moretti committed
499
    unless ($quiet=~ /on/i) { 
Sebastien Moretti's avatar
Sebastien Moretti committed
500
501
502
        print {*STDOUT} "
             Version:     $version
             Database:    $name_database
Sebastien Moretti's avatar
Sebastien Moretti committed
503
             Posted date: $posted\n\n";
Sebastien Moretti's avatar
Sebastien Moretti committed
504
    }
Sebastien Moretti's avatar
Sebastien Moretti committed
505
506
507
508
509
    shift (@list_pdb);  
    return (@list_pdb);
}

#-----------------------------------------------------------------------------------------------------------------------------
510
sub PARSING {    
Sebastien Moretti's avatar
Sebastien Moretti committed
511
    my ($list_pdb, $locale, $distant, $method, $quiet, $database, $gigablast) = @_;
Sebastien Moretti's avatar
Sebastien Moretti committed
512

513
514
515
516
    my (@list_pdb)        = @$list_pdb;
    my (@result_not_sort) = ();
    my $n = 0;
    open (my $SOR, '>', 'webblast.log') or die;
Sebastien Moretti's avatar
Sebastien Moretti committed
517

518
519
520
    if ( $gigablast=~ /^yes$/i ) { $locale = 2;$distant = 0;}
    if ( $gigablast=~ /^no$/i )  { $locale = 1;}
    if ( $distant==1 )           { $locale = 0;}
Sebastien Moretti's avatar
Sebastien Moretti committed
521

522
523
524
525
526
527
    foreach my $pdb_result(@list_pdb){
        my $query, my $length_query, my($pdb_id), my $comp=0;
        if ( $pdb_result=~/No hits found/m ){
            print {$SOR} $pdb_result;
            next;
        }
528

529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
        $pdb_result =~ s/ALIGNMENTS//;
        local $/ = undef;
        my (@intra_res) = split(/(?=\n\n>)/s, $pdb_result);


        if ( $distant==1 ){
            my $version_d, my $database_d, my $poste_d;
            undef $/; ($query, $length_query) = ( $intra_res[0] =~ /Query=\s+(\S+)\s+Length=\s*(\d+)/smo );
            $/ = "\n";
            open (F3, '<', 'web_tempo.result') or die ;
            while ($_=<F3>){
                if ($_=~ /(BLASTP\s+\S+)/o)        { $version_d  = $1;}
                if ($_=~ /Database:\s+(.+?)$/o)    { $database_d = $1;}
                if ($_=~ /Posted date:\s*(.+?)$/o) { $poste_d    = $1; last;}
            }
            close F3;
#           $database_d = $database if ( $database_d =~ m{/} );
            unless ($quiet=~ /on/i || $n>0){
                ++$n;
		        print {*STDOUT} "
Sebastien Moretti's avatar
Sebastien Moretti committed
549
550
             Version:     $version_d
             Database:    $database_d
Sebastien Moretti's avatar
Sebastien Moretti committed
551
             Posted date: $poste_d\n\n";
552
553
554
555
556
            }
        }
        else{
            ($query, $length_query) = ( $intra_res[0] =~ /\s*(.+?)\s.+?\(([\d,]+) letters/smo );
        }
557

558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
        shift(@intra_res) if ( exists($intra_res[1]) );

        foreach my $intra_res(@intra_res){ #look for the different results of the query
            my($aln_length, $identity) = ( $intra_res=~ /^\sIdentities = \d+\/(\d+)\s\((.+?)\)/im );
            my($recouvrement, $gap)    = &RECOVER($intra_res, $aln_length, $length_query);
            my($evalue)                = ( $intra_res=~ /Expect = (.+?)\s/im );
            my($bits)                  = ( $intra_res=~ /Score =\s+([\d.]+)\s/im );

            unless ( $method !~ /^geneid$/i ){
                if ( $comp<=$bits ){
                    $comp=$bits;
                }
                else{
                    last;
                }
            }

            if ( $query eq '' || $length_query eq '' || $aln_length eq '' || $identity eq '' || $recouvrement eq '' || $gap eq '' ){
                print {$SOR} " can't parse $pdb_result";
                next;
            }

            if ( $method =~ /^pdbid$/i ){
                if ( $locale==1 ){
                    ($pdb_id) = ( $intra_res=~ /^>(.{6})/im );
                    $pdb_id   =~ s/_//;
                    $pdb_id   = uc($pdb_id);
                }
                else{
                    ($pdb_id) = ( $intra_res=~ /^>pdb\|(.{6})/im );
                     $pdb_id  =~ s/\|//;
                }
                ($evalue) = ( $intra_res=~ /Expect = (.+?)\s/im );

                push (@result_not_sort, "$query\t$pdb_id\t$evalue\t$identity\t$recouvrement\t");
            }
            elsif ( $method =~/^geneid$/i ){
                if ( $database !~ /pdb/i && $database !~ /swiss/i && $locale =~ /1|2/){
                    while ( $intra_res=~ />.*?(gb|prf|emb|sp|pir|tpe|ref|prf|dbj|ddbj|pdb)[\|]+([A-Za-z0-9_\.]+?)(\s|\|(.{1}))/sg ){
                        my $databank = $1;
                        my $last     = $4;
                        my $refseq   = $2;
                        if ( $databank eq 'pdb' ){
                            $refseq .= $last;
                        }
                        $refseq =~ s/\.\d+$//;
                        push (@result_not_sort, "$query\t$refseq\t$identity\t$recouvrement\t$bits\t$evalue\t$databank");
                    }
                }
                elsif ( $database=~ /pdb|pdbaa/i && ($locale==1 || $locale==2) ){
                    my $refseq;
                    if ( $locale==1 ){
                        ($refseq) = ( $intra_res=~ />(.*?)\s/o );
                        $refseq   =~ s/_//;
                    }
                    else{
                        ($refseq) = ( $intra_res=~ /^>pdb\|(.{6})/im );
                        $refseq   =~ s/\|//;
                    }

                    unless ($refseq){
                        print {$SOR} $intra_res;
                        next;
                    }
                    push (@result_not_sort, "$query\t$refseq\t$identity\t$recouvrement\t$bits\t$evalue\tpdb");
                }
                elsif ( $distant==1 ){
                    my ($resul) = &MULTI_EQUIVALENT($query, $identity, $recouvrement, $bits, $evalue, $intra_res);
                    push (@result_not_sort, $resul);
                }
                elsif ( $database=~ /swiss/i ){
                    my ($refseq) = ( $intra_res=~ />.*?sp\|(.+?)\|/o );
                    unless ($refseq){
                        print {$SOR} $pdb_result;
                        next;
                    }
                    $refseq =~ s/\.\d+$//;
                    push (@result_not_sort, "$query\t$refseq\t$identity\t$recouvrement\t$bits\tswiss_prot");
                }
            }
            else {die;}
        }
640
    }
Sebastien Moretti's avatar
Sebastien Moretti committed
641
    close $SOR;
Sebastien Moretti's avatar
Sebastien Moretti committed
642
    undef (@list_pdb);
643

644
645
646
647
648
649
650
651
652
653
    if ( $method =~/^geneid$/i ){
        return (@result_not_sort);
    }
    else{
        my(@result_sort) = map  { $_->[1] }
                           sort { $b->[0]<=>$a->[0] }
                           map  { [/\t([\d.]+)%/,$_] }
                           @result_not_sort;
        undef(@result_not_sort);
        return (@result_sort);
654
    }
Sebastien Moretti's avatar
Sebastien Moretti committed
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
}

#-------------------------------------------------------------------------------------------------------------------

sub MULTI_EQUIVALENT
{
    my($query,$identity,$recouvrement,$bits,$evalue,$intra_res)=@_;

    my @result=();
    while  ($intra_res=~ />.*?(gb|prf|emb|sp|pir|tpe|ref|prf|dbj|ddbj|pdb)[\|]+([A-Za-z0-9_\.]+?)(\s|\|(.{1}))/g) 
    { 
	my $databank =$1;
	my $last     =$4;
	my $refseq   =$2;
	
670
	if ($databank eq 'pdb') { $refseq.=$last } 
Sebastien Moretti's avatar
Sebastien Moretti committed
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
	$refseq=~ s/\.\d+$//;
	push (@result,"$query\t$refseq\t$identity\t$recouvrement\t$bits\t$evalue\t$databank");
    }
    return (@result);
}
#--------------------------------------------------------------------------------------------------------------------
sub AFFICHAGE_REFSEQ_PARSING
{
    my($result_sort,$cover_tresh,$identity_treshold,$out_file)=@_;
    my(@result_sort)=@$result_sort, my(@name_gid)=();my@resultats=();my $afficher="";
   
(my($entete)= sprintf("%-40s %-25s %-10s %-12s %-10s %-10s %-10s","Sequence Name","Accession number","Databank","%Identity","%Cover","BITS","Evalue")); 
    
    foreach my $result_sort(@result_sort)
    {     
686
687
	my($seq_name,$refseq_name,$identiq,$cover,$bits,$evalue,$bank)= split("\t",$result_sort);
    $evalue =~ s/,$//; #To remove an additional comment with new blast release (2.2.17)
Sebastien Moretti's avatar
Sebastien Moretti committed
688
689
690
691
692
693
694
695
696
697
698
699
700
701
	($identiq)= split(/%/,$identiq);
	
	if ($identiq >= $identity_treshold && $cover >= $cover_tresh)
	{
	    push (@name_gid,">$seq_name\@$bank\_\_$refseq_name\n");
	    (($afficher).=  sprintf("%-40s %-25s %-10s %-12s %-10s %-10s %-10s ",$seq_name,$refseq_name,$bank,$identiq,$cover,$bits,$evalue));
	    $afficher.="\n";
	} 
	else {next;}	
    }

if ($afficher) { print "\n$entete\n\n"; print $afficher; }


702
703
704
705
if (@name_gid) { print {*STDOUT} "\n**********************************************************************\n\n"; }
if ($out_file) { open (SOR,">$out_file") or die "can not open $out_file"; print SOR @name_gid; }
print {*STDOUT} "\n", @name_gid;
close SOR;
Sebastien Moretti's avatar
Sebastien Moretti committed
706
707
708
709
710
}
#-------------------------------------------------------------------------------------------------------------

sub AFFICHAGE_PDB_PARSING 
{
711

Sebastien Moretti's avatar
Sebastien Moretti committed
712
713
    my($result_sort,$cover_tresh,$identity_treshold,$out_file)=@_;
    my(@result_sort)=@$result_sort, my @sortie=();
714
715
    
    print {*STDOUT} "\n\n",(my($en_tete)= sprintf("%-40s %-10s %-10s %-12s %-10s","Sequence Name","PDB_id","Evalue","Identity(%)","Cover(%)")),"\n\n"; 
Sebastien Moretti's avatar
Sebastien Moretti committed
716
717
718
719
720
721
722
723
724
    
    foreach my $result_sort(@result_sort)
    {     
	my($seq_name,$pdb_name,$EValue,$identiq,$cover)= split("\t",$result_sort);  	    
	($identiq)= split(/%/,$identiq);
	
	if ($identiq >= $identity_treshold && $cover >= $cover_tresh)
	{
	    push (@pdb_list,$pdb_name);
725
726
        $EValue =~ s/,$//;
	    print {*STDOUT} ((my $afficher)= sprintf("%-40s %-10s %-10s %-12s %-10s",$seq_name,$pdb_name,$EValue,$identiq,$cover)),"\n";
Sebastien Moretti's avatar
Sebastien Moretti committed
727
728
	    push (@sortie,">$seq_name _P_ $pdb_name\n");
	} 
729
	else {next;}		
Sebastien Moretti's avatar
Sebastien Moretti committed
730
731
    }
    undef(@result_sort);
732
    print {*STDOUT} "\n**********************************************************************\n\n";
Sebastien Moretti's avatar
Sebastien Moretti committed
733
734
735

#-- OUTFILE /STDOUT
    if   ($out_file) { open (SOR,">$out_file") or die "can not open $out_file"; print SOR @sortie; }
736
    print {*STDOUT} @sortie;
Sebastien Moretti's avatar
Sebastien Moretti committed
737
738
739
740
741
742
    close SOR;

}
#-----------------------------------------------------------------------------------------------------------------------------------
sub PROFILE
{
743
744
745
    my($list_pdb,$out_file,$distant)=@_;
    my(@list_pdb)=@$list_pdb,  my(@sortie)=();
    my %names=();   my $i=0;    my($name)='';
Sebastien Moretti's avatar
Sebastien Moretti committed
746
747
748

    open (SOR1,">$out_file") or die;
    foreach my $pdb_result(@list_pdb)
749
    {     	
Sebastien Moretti's avatar
Sebastien Moretti committed
750
751
	if ($pdb_result =~ /No hits found/i) { next; }
	else
752
753
754
755
	{
	    ++$i;
	    if ($distant==1) {($name)   =($pdb_result =~ /Query=\s*(.+?)Length/smoi)  or die "\nparse error in distant profile\n";}
	    else             {($name)   =($pdb_result =~ /\s*(.+?)\(.*?letters/ismo)  or die "\nparse error in profile\n";}
Sebastien Moretti's avatar
Sebastien Moretti committed
756
	  
757
758
	    my($name1)= ($name=~ /(.+?)\s+$/);
 
Sebastien Moretti's avatar
Sebastien Moretti committed
759
	    open (SOR,">tempo_file_profile") or die "can not open tempo_file_profile";
760
	    print SOR "Query= $pdb_result";
Sebastien Moretti's avatar
Sebastien Moretti committed
761
762
	    close SOR;	    
	    
763
	    open(COM,"|t_coffee -other_pg seq_reformat -input blast_aln -in tempo_file_profile -output fasta_aln -out ${i}.profile");	     
Sebastien Moretti's avatar
Sebastien Moretti committed
764
	    close COM;
765
	    push (@sortie,">$name1 _R_ ${i}.profile\n");
Sebastien Moretti's avatar
Sebastien Moretti committed
766
767
768
769
770
771
     
	} 	
    }    
    unlink("tempo_file_profile");
    undef(@list_pdb); 
  
772
    print {*STDERR} "\n**********************************************************************\n\n";
Sebastien Moretti's avatar
Sebastien Moretti committed
773
774
#-- OUTFILE /STDOUT
    if   ($out_file) { open (SOR1,">$out_file") or die "can not open $out_file"; print SOR1 @sortie; }
775
    print {*STDOUT} @sortie;
Sebastien Moretti's avatar
Sebastien Moretti committed
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
    close SOR1;
         
}

#--------------------------------------------------------------------------------------------------------------------------------
sub ORGN
{   
    my($organism)=@_;
    $organism=~ s/_/ /;

    my(%orgs)= (
		
		'Homo sapiens'         =>'1',
		'Bos taurus'             =>'1',
		'Gallus gallus'         =>'1',
		'Viruses'              =>'1',
		'Bacteria'            =>'1',           
		'Eukaryota'            =>'1',
		'Mammalia'              =>'1',
		'Vertebrata'              =>'1',
		'All organisms'          =>'1',
		'Fungi'                 =>'1',
		'Primates'             =>'1',
		'Archaea'               =>'1',
                'Arabidopsis thaliana'   =>'1',
                'Caenorhabditis elegans'  =>'1',
                'Escherichia coli'        =>'1',
		'Mus musculus'             =>'1',
                'Drosophila melanogaster'   =>'1',
		);

    if (exists $orgs{$organism}) 
    { $organism=~ s/ /+/g; return ($organism); }
809
    else { print {*STDERR} "organism not valid or syntax error, replace space by \"_\" \n"; &HELP(); }
Sebastien Moretti's avatar
Sebastien Moretti committed
810
811
812
813
814
815
   
} 
#------------------------------------------------------------------------------------------------------------------------------------

sub LIST_ORGA
{
Sebastien Moretti's avatar
Sebastien Moretti committed
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
    my (%orgs) = (

        'Homo sapiens'            =>'1',
        'Bos taurus'              =>'1',
        'Gallus gallus'           =>'1',
        'Viruses'                 =>'1',
        'Bacteria'                =>'1',           
        'Eukaryota'               =>'1',
        'Mammalia'                =>'1',
        'Vertebrata'              =>'1',
        'All organisms'           =>'1',
        'Fungi'                   =>'1',
        'Primates'                =>'1',
        'Archaea'                 =>'1',
        'Arabidopsis thaliana'    =>'1',
        'Caenorhabditis elegans'  =>'1',
        'Escherichia coli'        =>'1',
        'Mus musculus'            =>'1',
        'Drosophila melanogaster' =>'1',
Sebastien Moretti's avatar
Sebastien Moretti committed
835
836
		);
    
Sebastien Moretti's avatar
Sebastien Moretti committed
837
838
839
840
    my (@cle) = keys(%orgs);
    foreach my $cle(@cle){
        $cle=~ s/ /_/g;
    }
Sebastien Moretti's avatar
Sebastien Moretti committed
841
842
    return (@cle);
}
843