SliceAdaptor.pm 44.9 KB
Newer Older
1 2

#
Simon Potter's avatar
pod  
Simon Potter committed
3
# Ensembl module for Bio::EnsEMBL::DBSQL::SliceAdaptor
4 5 6 7 8 9 10 11 12 13 14
#
# Cared for by Ewan Birney <ensembl-dev@ebi.ac.uk>
#
# Copyright Ewan Birney
#
# You may distribute this module under the same terms as perl itself

# POD documentation - main docs before the code

=head1 NAME

15 16
Bio::EnsEMBL::DBSQL::SliceAdaptor - A database aware adaptor responsible for
the creation of Slice objects.
17 18 19

=head1 SYNOPSIS

20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
  my $db = Bio::EnsEMBL::DBSQL::DBAdaptor->new(...);

  my $slice_adaptor = $db->get_SliceAdaptor();

  #get a slice on the entire chromosome X
  my $chr_slice = $slice_adaptor->fetch_by_region('chromosome','X');

  #get a slice for each clone in the database
  foreach my $cln_slice (@{$slice_adaptor->fetch_all('clone')}) {
    #do something with clone
  }

  #get a slice which is part of NT_004321
  my $spctg_slice = $slice_adaptor->fetch_by_region('supercontig','NT_004321',
                                                    200_000, 600_000);
35 36 37 38


=head1 DESCRIPTION

39 40 41
This module is responsible for fetching Slices representing genomic regions
from a database.  Details on how slices can be used are in the
Bio::EnsEMBL::Slice module.
42 43 44 45 46

=head1 AUTHOR - Ewan Birney

This modules is part of the Ensembl project http://www.ensembl.org

47
=head1 CONTACT
48

49
Email ensembl-dev@ebi.ac.uk
50

51
=head1 METHODS
52 53 54 55

=cut


56
package Bio::EnsEMBL::DBSQL::SliceAdaptor;
57 58 59
use vars qw(@ISA);
use strict;

60

61 62
use Bio::EnsEMBL::DBSQL::BaseAdaptor;
use Bio::EnsEMBL::Slice;
63
use Bio::EnsEMBL::DBSQL::DBAdaptor;
64
use Bio::EnsEMBL::Mapper;
65

66 67
use POSIX qw(ceil floor);

68 69
use Bio::EnsEMBL::Utils::Exception qw(throw deprecate warning);
use Bio::EnsEMBL::Utils::Cache; #CPAN LRU cache
70

71
@ISA = ('Bio::EnsEMBL::DBSQL::BaseAdaptor');
72

73
my $SEQ_REGION_CACHE_SIZE = 1000;
74

75 76
sub new {
  my $caller = shift;
77

78
  my $class = ref($caller) || $caller;
79

80
  my $self = $class->SUPER::new(@_);
81

82 83
  my %name_cache;
  my %id_cache;
84

85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103
  tie(%name_cache, 'Bio::EnsEMBL::Utils::Cache', $SEQ_REGION_CACHE_SIZE);
  tie(%id_cache, 'Bio::EnsEMBL::Utils::Cache', $SEQ_REGION_CACHE_SIZE);

  $self->{'_name_cache'} = \%name_cache;
  $self->{'_id_cache'} = \%id_cache;

  return $self;
}


=head2 fetch_by_region

  Arg [1]    : string $coord_system_name
               The name of the coordinate system of the slice to be created
               This may be a name of an acutal coordinate system or an alias
               to a coordinate system.  Valid aliases are 'seqlevel' or
               'toplevel'.
  Arg [2]    : string $seq_region_name
               The name of the sequence region that the slice will be
104
               created on.                
105 106 107 108 109 110 111 112
  Arg [3]    : int $start (optional, default = 1)
               The start of the slice on the sequence region
  Arg [4]    : int $end (optional, default = seq_region length)
               The end of the slice on the sequence region
  Arg [5]    : int $strand (optional, default = 1)
               The orientation of the slice on the sequence region
  Arg [6]    : string $version (optional, default = default version)
               The version of the coordinate system to use (e.g. NCBI33)
113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131
  Example    : $slice = $slice_adaptor->fetch_by_region('chromosome', 'X');
               $slice = $slice_adaptor->fetch_by_region('clone', 'AC008066.4');
  Description: Retrieves a slice on the requested region.  At a minimum the
               name of the coordinate system and the name of the seq_region to
               fetch must be provided.

               Some fuzzy matching is performed if no exact match for
               the provided name is found.  This allows clones to be
               fetched even when their version is not known.  For
               example fetch_by_region('clone', 'AC008066') will
               retrieve the sequence_region with name 'AC008066.4'.

               If the requested seq_region is not found in the database undef
               is returned.

  Returntype : Bio::EnsEMBL::Slice or undef
  Exceptions : throw if no seq_region_name is provided
               throw if invalid coord_system_name is provided
               throw if start > end is provided
132
  Caller     : general
133 134 135

=cut

136 137 138 139
sub fetch_by_region {
  my ($self, $coord_system_name, $seq_region_name,
      $start, $end, $strand, $version) = @_;

140 141 142 143
  $start  = 1 if(!defined($start));
  $strand = 1 if(!defined($strand));

  throw('seq_region_name argument is required') if(!defined($seq_region_name));
144 145 146 147

  my $csa = $self->db()->get_CoordSystemAdaptor();
  my $coord_system = $csa->fetch_by_name($coord_system_name,$version);

148 149 150 151 152 153 154 155 156
  if(!$coord_system) {
    throw("Unknown coordinate system:\n name='$coord_system_name' " .
          "version='$version'\n");
  }
  if($coord_system->is_top_level()) {
    throw("Cannot fetch_by_region using the 'toplevel' pseudo coordinate " .
          "system.\n");
  }

157 158 159 160 161 162 163 164 165 166 167 168 169 170
  #check the cache so we only go to the db if necessary
  my $name_cache = $self->{'_name_cache'};
  my $key = lc(join(':',$seq_region_name,
                    $coord_system->name(),
                    $coord_system->version));

  my $length;

  if(exists($name_cache->{$key})) {
    $length = $name_cache->{$key}->[1];
  } else {
    my $sth = $self->prepare("SELECT seq_region_id, length " .
                             "FROM seq_region " .
                             "WHERE name = ? AND coord_system_id = ?");
Laura Clarke's avatar
 
Laura Clarke committed
171
 
172 173 174
    #force seq_region_name cast to string so mysql cannot treat as int
    $sth->execute("$seq_region_name", $coord_system->dbID());

175 176
    if($sth->rows() == 0) {
      $sth->finish();
Laura Clarke's avatar
 
Laura Clarke committed
177
 
178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226
      #do fuzzy matching, assuming that we are just missing a version on 
      #the end of the seq_region name
   
      $sth = $self->prepare("SELECT name, seq_region_id, length " .
                            "FROM   seq_region " .
                            "WHERE  name LIKE ?" .
                            "AND    coord_system_id = ?");

      $sth->execute("$seq_region_name.%", $coord_system->dbID());

      my $prefix_len = length($seq_region_name) + 1;
      my $highest_version = undef;

      # find the fuzzy-matched seq_region with the highest postfix (which ought
      # to be a version)

      my ($tmp_name, $id, $tmp_length);

      while(($tmp_name, $id, $tmp_length) = $sth->fetchrow_array()) {
        $key = lc(join(':',$tmp_name,
                       $coord_system->name(),
                       $coord_system->version));
        $name_cache->{$key}         = [$id,$tmp_length];
        $self->{'_id_cache'}->{$id} = [$tmp_name,$tmp_length,$coord_system];
        
        my $version = substr($tmp_name, $prefix_len);

        #skip versions which are non-numeric and apparently not versions
        next if($version !~ /^\d+$/);

        if(!defined($highest_version) || ($version <=> $highest_version) > 0) {
          $seq_region_name = $tmp_name;
          $length          = $tmp_length;
          $highest_version = $version;
        } 
      } 
      $sth->finish();

      #return if we did not find any appropriate match:
      return undef if(!defined($highest_version));
    } else {

      my $id;
      ($id, $length) = $sth->fetchrow_array();
      $sth->finish();

      #cache results to speed up future queries
      $name_cache->{$key}         = [$id,$length];
      $self->{'_id_cache'}->{$id} = [$seq_region_name, $length, $coord_system];
227
    }
228
  }
229

230
  $end = $length if(!defined($end));
231
  
232 233 234 235
  if($end < $start) {
    throw('start [$start] must be less than or equal to end [$end]');
  }

236 237 238 239 240 241 242
  return Bio::EnsEMBL::Slice->new(-COORD_SYSTEM      => $coord_system,
                                  -SEQ_REGION_NAME   => $seq_region_name,
                                  -SEQ_REGION_LENGTH => $length,
                                  -START             => $start,
                                  -END               => $end,
                                  -STRAND            => $strand,
                                  -ADAPTOR           => $self);
243 244 245 246
}



247
=head2 fetch_by_name
248

Graham McVicker's avatar
Graham McVicker committed
249
  Arg [1]    : string $name
250 251 252 253 254 255 256 257 258
  Example    : $name  = 'chromosome:NCBI34:X:1000000:2000000:1';
               $slice = $slice_adaptor->fetch_by_name($name);
               $slice2 = $slice_adaptor->fetch_by_name($slice3->name());
  Description: Fetches a slice using a slice name (i.e. the value returned by
               the Slice::name method).  This is useful if you wish to 
               store a unique identifier for a slice in a file or database or
               pass a slice over a network.
               Slice::name allows you to serialise/marshall a slice and this
               method allows you to deserialise/unmarshal it.
259 260 261 262 263
                
               Returns undef if no seq_region with the provided name exists in
               the database.

  Returntype : Bio::EnsEMBL::Slice or undef
264 265
  Exceptions : throw if incorrent arg provided
  Caller     : Pipeline
266 267 268

=cut

269 270 271
sub fetch_by_name {
  my $self = shift;
  my $name = shift;
Graham McVicker's avatar
Graham McVicker committed
272

273 274
  if(!$name) {
    throw("name argument is required");
Graham McVicker's avatar
Graham McVicker committed
275
  }
276 277 278 279 280

  my @array = split(/:/,$name);

  if(@array != 6) {
    throw("Malformed slice name [$name].  Format is " .
Laura Clarke's avatar
 
Laura Clarke committed
281
        "coord_system:version:name:start:end:strand");
282 283 284 285 286
  }

  my ($cs_name, $cs_version, $seq_region, $start, $end, $strand) = @array;

  return $self->fetch_by_region($cs_name,$seq_region, $start,
287
                                $end, $strand, $cs_version);
Graham McVicker's avatar
Graham McVicker committed
288
}
Graham McVicker's avatar
Graham McVicker committed
289 290 291



292 293 294 295 296 297 298 299
=head2 fetch_by_seq_region_id

  Arg [1]    : string $seq_region_id
               The internal identifier of the seq_region to create this slice
               on
  Example    : $slice = $slice_adaptor->fetch_by_seq_region_id(34413);
  Description: Creates a slice object of an entire seq_region using the
               seq_region internal identifier to resolve the seq_region.
300 301
               Returns undef if no such slice exists.
  Returntype : Bio::EnsEMBL::Slice or undef
Graham McVicker's avatar
Graham McVicker committed
302 303
  Exceptions : none
  Caller     : general
Graham McVicker's avatar
Graham McVicker committed
304 305 306

=cut

307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322
sub fetch_by_seq_region_id {
  my ($self, $seq_region_id) = @_;

  my $id_cache = $self->{'_id_cache'};

  my ($name, $length, $cs);

  if(exists $id_cache->{$seq_region_id}) {
    ($name, $length, $cs) = @{$id_cache->{$seq_region_id}};
  } else {
    my $sth = $self->prepare("SELECT name, length, coord_system_id " .
                             "FROM seq_region " .
                             "WHERE seq_region_id = ?");

    $sth->execute($seq_region_id);

323
    return undef if($sth->rows() == 0);
324 325 326 327 328 329 330 331 332 333 334 335 336

    my $cs_id;
    ($name, $length, $cs_id) = $sth->fetchrow_array();
    $sth->finish();

    $cs = $self->db->get_CoordSystemAdaptor->fetch_by_dbID($cs_id);

    #cache results to speed up repeated queries
    $id_cache->{$seq_region_id} = [$name, $length, $cs];
    my $key = lc(join(':', $name, $cs->name, $cs->version));
    $self->{'_name_cache'}->{$key} = [$seq_region_id, $length];
  }

337 338 339 340 341 342 343
  return Bio::EnsEMBL::Slice->new(-COORD_SYSTEM      => $cs,
                                  -SEQ_REGION_NAME   => $name,
                                  -SEQ_REGION_LENGTH => $length,
                                  -START             => 1,
                                  -END               => $length,
                                  -STRAND            => 1,
                                  -ADAPTOR           => $self);
Arne Stabenau's avatar
Arne Stabenau committed
344 345
}

Graham McVicker's avatar
Graham McVicker committed
346

347 348

=head2 get_seq_region_id
Graham McVicker's avatar
Graham McVicker committed
349

Arne Stabenau's avatar
Arne Stabenau committed
350
  Arg [1]    : Bio::EnsEMBL::Slice $slice
351 352 353 354 355 356 357 358 359 360
               The slice to fetch a seq_region_id for
  Example    : $srid = $slice_adaptor->get_seq_region_id($slice);
  Description: Retrieves the seq_region id (in this database) given a slice
               Seq region ids are not stored on the slices themselves
               because they are intended to be somewhat database independant
               and seq_region_ids vary accross databases.
  Returntype : int
  Exceptions : throw if the seq_region of the slice is not in the db
               throw if incorrect arg provided
  Caller     : BaseFeatureAdaptor
Arne Stabenau's avatar
Arne Stabenau committed
361 362

=cut
Graham McVicker's avatar
Graham McVicker committed
363

364 365 366
sub get_seq_region_id {
  my $self = shift;
  my $slice = shift;
Graham McVicker's avatar
Graham McVicker committed
367

368 369 370 371 372 373 374 375 376
  if(!$slice || !ref($slice) || !$slice->isa('Bio::EnsEMBL::Slice')) {
    throw('Slice argument is required');
  }

  my $cs_name = $slice->coord_system->name();
  my $cs_version = $slice->coord_system->version();
  my $seq_region_name = $slice->seq_region_name();

  my $key = lc(join(':', $seq_region_name,$cs_name,$cs_version));
377

378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409
  my $name_cache = $self->{'_name_cache'};

  if(exists($name_cache->{$key})) {
    return $name_cache->{$key}->[0];
  }

  my $csa = $self->db()->get_CoordSystemAdaptor();
  my $coord_system = $csa->fetch_by_name($cs_name,$cs_version);

  my $sth = $self->prepare("SELECT seq_region_id, length " .
                           "FROM seq_region " .
                           "WHERE name = ? AND coord_system_id = ?");

  #force seq_region_name cast to string so mysql cannot treat as int
  $sth->execute("$seq_region_name", $coord_system->dbID());

  if($sth->rows() != 1) {
    throw("Non existant or ambigous seq_region:\n" .
          "  coord_system=[$cs_name],\n" .
          "  name=[$seq_region_name],\n" .
          "  version=[$cs_version]");
  }

  my($seq_region_id, $length) = $sth->fetchrow_array();
  $sth->finish();

  #cache information for future requests
  $name_cache->{$key} = [$seq_region_id, $length];
  $self->{'_id_cache'}->{$seq_region_id} =
    [$seq_region_name, $length, $coord_system];

  return $seq_region_id;
Arne Stabenau's avatar
Arne Stabenau committed
410
}
411 412


413
=head2 get_seq_region_attribs
Web Admin's avatar
Web Admin committed
414

415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437
  Arg [1]    : Bio::EnsEMBL::Slice $slice
               The slice to fetch attributes for
  Example    : %attribs = %{$slice_adaptor->get_seq_region_attribs($slice)};
               ($htg_phase) = @{$attribs->{'htg_phase'} || []};
               @synonyms    = @{$attribs->{'synonym'} || []};
  Description: Retrieves a reference to a hash containing attrib code values
               and listref value keys.
  Returntype : hashref
  Exceptions : throw if the seq_region of the slice is not in the db
               throw if incorrect arg provided
  Caller     : Bio::EnsEMBL::Slice

=cut

sub get_seq_region_attribs {
  my $self = shift;
  my $slice = shift;

  if(!ref($slice) || !$slice->isa('Bio::EnsEMBL::Slice')) {
    throw('Slice argument is required');
  }

  my $srid = $self->get_seq_region_id($slice);
Web Admin's avatar
Web Admin committed
438

439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478
  if(!$srid) {
    throw('Slice is not on a seq_region stored in this database.');
  }

  $self->{'_attribs_cache'} ||= {};
  if($self->{'_attribs_cache'}->{$srid}) {
    return $self->{'_attribs_cache'}->{$srid};
  }

  my $sth = $self->prepare('SELECT at.code, sra.value ' .
                           'FROM   seq_region_attrib sra, attrib_type at ' .
                           'WHERE  sra.seq_region_id = ? ' .
                           'AND    at.attrib_type_id = sra.attrib_type_id');

  $sth->execute($srid);

  my($code, $attrib);
  $sth->bind_columns(\$code, \$attrib);

  my %attrib_hash;
  while($sth->fetch()) {
    $attrib_hash{$code} ||= [];
    push @{$attrib_hash{$code}}, $attrib;
  }

  $sth->finish();
  $self->{'_attribs_cache'} = \%attrib_hash;
  return \%attrib_hash;
}


=head2 fetch_all

  Arg [1]    : string $coord_system_name
               The name of the coordinate system to retrieve slices of.
               This may be a name of an acutal coordinate system or an alias
               to a coordinate system.  Valid aliases are 'seqlevel' or
               'toplevel'.
  Arg [2]    : string $coord_system_version (optional)
               The version of the coordinate system to retrieve slices of
479
  Arg [3]    : bool $include_non_reference (optional)
480 481 482
               If this argument is not provided then only reference slices
               will be returned. If set, both reference and non refeference
               slices will be rerurned.
483
  Arg [4]    : int $no_duplicates (optional)
484
               If set no duplicate entries will be returned.
485

486 487 488 489 490 491 492 493 494
  Example    : @chromos = @{$slice_adaptor->fetch_all('chromosome','NCBI33')};
               @contigs = @{$slice_adaptor->fetch_all('contig')};


  Description: Retrieves slices of all seq_regions for a given coordinate
               system.  This is analagous to the methods fetch_all which were
               formerly on the ChromosomeAdaptor, RawContigAdaptor and
               CloneAdaptor classes.  Slices fetched span the entire
               seq_regions and are on the forward strand.
495 496
               If the coordinate system with the provided name and version
               does not exist an empty list is returned.
497
               If the coordinate system name provided is 'toplevel', all
498
               non-redundant toplevel slices are returned.
499 500 501 502

               Retrieved slices can be broken into smaller slices using the
               Bio::EnsEMBL::Utils::Slice module.

503
  Returntype : listref of Bio::EnsEMBL::Slices
504
  Exceptions : none
505
  Caller     : general
Web Admin's avatar
Web Admin committed
506 507 508

=cut

509 510 511 512 513
sub fetch_all {
  my $self = shift;
  my $cs_name = shift;
  my $cs_version = shift || '';

514
  my ($include_non_reference, $include_duplicates) = @_;
Web Admin's avatar
Web Admin committed
515

516 517 518
  #
  # verify existance of requested coord system and get its id
  #
519 520 521 522 523 524
  my $csa       = $self->db->get_CoordSystemAdaptor();
  my $orig_cs   = $csa->fetch_by_name($cs_name, $cs_version);

  return [] if(!$orig_cs);

  my $sth;
525

526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544
  my %bad_vals=();
  #
  # Get a hash of non reference seq regions
  #
  if(!$include_non_reference){
    my $sth2 = $self->prepare(
			      "SELECT sr.seq_region_id ".
			      "FROM seq_region sr, seq_region_attrib sra, attrib_type at ".
			      " WHERE at.code='non_ref'".
			      "  AND sra.seq_region_id=sr.seq_region_id ".
			      "  AND at.attrib_type_id=sra.attrib_type_id " );
    $sth2->execute();
    my ($seq_region_id);
    $sth2->bind_columns(\$seq_region_id);
    while($sth2->fetch()) {
      $bad_vals{$seq_region_id} = 1;
    }
    $sth2->finish();
  }
545
  #
546
  # Retrieve the seq_regions from the database
547
  #
548
  if($orig_cs->is_top_level()) {
549 550 551 552 553 554 555 556
   $sth =
       $self->prepare("SELECT sr.seq_region_id, sr.name, sr.length, " .
                      "       sr.coord_system_id " .
                      "FROM seq_region sr, " .
                      "     seq_region_attrib sra, attrib_type at " .
                      "WHERE at.code='toplevel' " .
                      "AND   at.attrib_type_id=sra.attrib_type_id " .
                      "AND   sra.seq_region_id=sr.seq_region_id");
557 558
    $sth->execute();
  } else {
559 560 561 562 563
     $sth =
       $self->prepare('SELECT seq_region_id, name, length, coord_system_id ' .
                      'FROM   seq_region ' .
                      'WHERE  coord_system_id =?');
     $sth->execute($orig_cs->dbID);
564
  }
565

566 567
  my ($seq_region_id, $name, $length, $cs_id);
  $sth->bind_columns(\$seq_region_id, \$name, \$length, \$cs_id);
568 569 570

  my $name_cache = $self->{'_name_cache'};
  my $id_cache   = $self->{'_id_cache'};
571
  my $cache_count = 0;
572 573 574

  my @out;
  while($sth->fetch()) {
575 576
    if(!defined($bad_vals{$seq_region_id})){
      my $cs = $csa->fetch_by_dbID($cs_id);
577

578
      if(!$cs) {
579
        throw("seq_region $name references non-existent coord_system $cs_id.");
580
      }
581

582
      my $cs_key = lc($cs->name().':'.$cs_version);
583

584 585 586
      #cache values for future reference, but stop adding to the cache once we
      #we know we have filled it up
      if($cache_count < $SEQ_REGION_CACHE_SIZE) {
587 588 589 590
        my $key = lc($name) . ':'. $cs_key;
        $name_cache->{$key} = [$seq_region_id, $length];
        $id_cache->{$seq_region_id} = [$name, $length, $cs];
        $cache_count++;
591
      }
592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609

      my $slice = Bio::EnsEMBL::Slice->new
        (-START             => 1,
         -END               => $length,
         -STRAND            => 1,
         -SEQ_REGION_NAME   => $name,
         -SEQ_REGION_LENGTH => $length,
         -COORD_SYSTEM      => $cs,
         -ADAPTOR           => $self);

      if(!defined($include_duplicates) or !$include_duplicates){
        #do not include duplicates
        my @dup = @{$self->fetch_normalized_slice_projection($slice)};
        foreach my $dup_test( @dup){
          if($dup_test->[2]->get_seq_region_id == $slice->get_seq_region_id){
            push @out, $dup_test->[2];
          }
        }
610
      }
611 612
      else {
        push @out, $slice;
613
      }
614
    }
615
  }
616
  $sth->finish();
617

618
  return \@out;
Web Admin's avatar
Web Admin committed
619 620 621
}


622 623 624 625
sub deleteObj {
  my $self = shift;

  $self->SUPER::deleteObj;
626

627 628 629 630 631 632 633
  $self->{'_id_cache'} = {};
  $self->{'_name_cache'} = {};
  $self->{'_exc_cache'} = {};
}



Web Admin's avatar
Web Admin committed
634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663
=head2 fetch_by_band

 Title   : fetch_by_band
 Usage   :
 Function: create a Slice representing a series of bands
 Example :
 Returns :
 Args    : the band name

=cut

sub fetch_by_band {
  my ($self,$band) = @_;

  my $sth = $self->db->prepare
        ("select s.name,max(k.seq_region_id)-min(k.seq_region_id, min(k.seq_region_start), max(k.seq_region_id) " .
         "from karyotype as k " .
         "where k.band like ? and k.seq_region_id = s.seq_region_id");

  $sth->execute( "$band%" );
  my ( $seq_region_name, $discrepancy, $seq_region_start, $seq_region_end) = $sth->fetchrow_array;

  if($seq_region_name && $discrepancy>0) {
    throw("Band maps to multiple seq_regions");
  } else {
    return $self->fetch_by_region('toplevel',$seq_region_name,$seq_region_start,$seq_region_end);
  }
  throw("Band not recognised in database");
}

664 665 666 667 668 669 670 671 672 673 674 675 676 677
=head2 fetch_by_chr_band

 Title   : fetch_by_chr_band
 Usage   :
 Function: create a Slice representing a series of bands
 Example :
 Returns :
 Args    : the band name

=cut

sub fetch_by_chr_band {
  my ($self,$chr,$band) = @_;

Web Admin's avatar
Web Admin committed
678
  my $chr_slice = $self->fetch_by_region('toplevel', $chr);
679 680 681 682

  my $seq_region_id = $self->get_seq_region_id($chr_slice);

  my $sth = $self->db->prepare
Web Admin's avatar
Web Admin committed
683
        ("select min(k.seq_region_start), max(k.seq_region_end) " .
684 685 686 687 688 689 690
         "from karyotype as k " .
         "where k.seq_region_id = ? and k.band like ?");

  $sth->execute( $seq_region_id, "$band%" );
  my ( $slice_start, $slice_end) = $sth->fetchrow_array;

  if(defined $slice_start) {
Web Admin's avatar
Web Admin committed
691
    return $self->fetch_by_region('toplevel',$chr,$slice_start,$slice_end);
692 693 694
  }

  throw("Band not recognised in database");
695 696 697 698
}



Web Admin's avatar
Web Admin committed
699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730
=head2 fetch_by_exon_stable_id

  Arg [1]    : string $exonid
               The stable id of the exon around which the slice is 
               desired
  Arg [2]    : (optional) int $size
               The length of the flanking regions the slice should encompass 
               on either side of the exon (0 by default)
  Example    : $slc = $sa->fetch_by_exon_stable_id('ENSE00000302930',10);
  Description: Creates a slice around the region of the specified exon. 
               If a context size is given, the slice is extended by that 
               number of basepairs on either side of the
               transcript.
  Returntype : Bio::EnsEMBL::Slice
  Exceptions : Thrown if the exon is not in the database.
  Caller     : general

=cut

sub fetch_by_exon_stable_id{
  my ($self,$exonid,$size) = @_;

  throw('Exon argument is required.') if(!$exonid);

  my $ea = $self->db->get_ExonAdaptor;
  my $exon = $ea->fetch_by_stable_id($exonid);

  throw("Exon [$exonid] does not exist in DB.") if(!$exon);

  return $self->fetch_by_Feature($exon, $size);
}

Graham McVicker's avatar
Graham McVicker committed
731
=head2 fetch_by_transcript_stable_id
732

Graham McVicker's avatar
Graham McVicker committed
733 734 735 736 737 738 739 740 741
  Arg [1]    : string $transcriptid
               The stable id of the transcript around which the slice is 
               desired
  Arg [2]    : (optional) int $size
               The length of the flanking regions the slice should encompass 
               on either side of the transcript (0 by default)
  Example    : $slc = $sa->fetch_by_transcript_stable_id('ENST00000302930',10);
  Description: Creates a slice around the region of the specified transcript. 
               If a context size is given, the slice is extended by that 
742 743
               number of basepairs on either side of the
               transcript.
Graham McVicker's avatar
Graham McVicker committed
744
  Returntype : Bio::EnsEMBL::Slice
745
  Exceptions : Thrown if the transcript is not in the database.
Graham McVicker's avatar
Graham McVicker committed
746
  Caller     : general
747 748 749

=cut

Graham McVicker's avatar
Graham McVicker committed
750
sub fetch_by_transcript_stable_id{
751 752
  my ($self,$transcriptid,$size) = @_;

753 754
  throw('Transcript argument is required.') if(!$transcriptid);

755
  my $ta = $self->db->get_TranscriptAdaptor;
756
  my $transcript = $ta->fetch_by_stable_id($transcriptid);
757

758 759 760 761
  throw("Transcript [$transcriptid] does not exist in DB.") if(!$transcript);

  return $self->fetch_by_Feature($transcript, $size);
}
762

Graham McVicker's avatar
Graham McVicker committed
763 764


Graham McVicker's avatar
Graham McVicker committed
765 766
=head2 fetch_by_transcript_id

Graham McVicker's avatar
Graham McVicker committed
767 768 769 770 771 772 773
  Arg [1]    : int $transcriptid
               The unique database identifier of the transcript around which 
               the slice is desired
  Arg [2]    : (optional) int $size
               The length of the flanking regions the slice should encompass 
               on either side of the transcript (0 by default)
  Example    : $slc = $sa->fetch_by_transcript_id(24, 1000);
774 775 776 777
  Description: Creates a slice around the region of the specified transcript.
               If a context size is given, the slice is extended by that
               number of basepairs on either side of the
               transcript.
Graham McVicker's avatar
Graham McVicker committed
778
  Returntype : Bio::EnsEMBL::Slice
779 780
  Exceptions : throw on incorrect args
               throw if transcript is not in database
Graham McVicker's avatar
Graham McVicker committed
781
  Caller     : general
782 783 784

=cut

Graham McVicker's avatar
Graham McVicker committed
785
sub fetch_by_transcript_id {
786
  my ($self,$transcriptid,$size) = @_;
Graham McVicker's avatar
Graham McVicker committed
787

788
  throw('Transcript id argument is required.') if(!$transcriptid);
Graham McVicker's avatar
Graham McVicker committed
789

790 791 792 793 794 795
  my $transcript_adaptor = $self->db()->get_TranscriptAdaptor();
  my $transcript = $transcript_adaptor->fetch_by_dbID($transcriptid);

  throw("Transcript [$transcriptid] does not exist in DB.") if(!$transcript);

  return $self->fetch_by_Feature($transcript, $size);
796 797
}

798 799


Graham McVicker's avatar
Graham McVicker committed
800
=head2 fetch_by_gene_stable_id
801

Graham McVicker's avatar
Graham McVicker committed
802
  Arg [1]    : string $geneid
803
               The stable id of the gene around which the slice is
Graham McVicker's avatar
Graham McVicker committed
804 805
               desired
  Arg [2]    : (optional) int $size
Graham McVicker's avatar
Graham McVicker committed
806
               The length of the flanking regions the slice should encompass
Graham McVicker's avatar
Graham McVicker committed
807 808
               on either side of the gene (0 by default)
  Example    : $slc = $sa->fetch_by_transcript_stable_id('ENSG00000012123',10);
Graham McVicker's avatar
Graham McVicker committed
809 810 811
  Description: Creates a slice around the region of the specified gene.
               If a context size is given, the slice is extended by that
               number of basepairs on either side of the gene.
812
               The slice will be created in the genes native coordinate system.
Graham McVicker's avatar
Graham McVicker committed
813
  Returntype : Bio::EnsEMBL::Slice
814 815
  Exceptions : throw on incorrect args
               throw if transcript does not exist
Graham McVicker's avatar
Graham McVicker committed
816
  Caller     : general
817 818 819

=cut

820 821
sub fetch_by_gene_stable_id {
  my ($self,$geneid,$size) = @_;
822

823
  throw('Gene argument is required.') if(!$geneid);
824

825 826
  my $gene_adaptor = $self->db->get_GeneAdaptor();
  my $gene = $gene_adaptor->fetch_by_stable_id($geneid);
827

828
  throw("Gene [$geneid] does not exist in DB.") if(!$gene);
829

830
  return $self->fetch_by_Feature($gene, $size);
831 832 833
}


834

835
=head2 fetch_by_Feature
Graham McVicker's avatar
Graham McVicker committed
836

837 838 839 840
  Arg [1]    : Bio::EnsEMBL::Feature $feat
               The feature to fetch the slice around
  Arg [2]    : int size (optional)
               The desired number of flanking basepairs around the feature.
841 842
               The size may also be provided as a percentage of the feature 
               size such as 200% or 80.5%.
843 844 845 846 847 848
  Example    : $slice = $slice_adaptor->fetch_by_Feature($feat, 100);
  Description: Retrieves a slice around a specific feature.  All this really
               does is return a resized version of the slice that the feature
               is already on. Note that slices returned from this method
               are always on the forward strand of the seq_region regardless of
               the strandedness of the feature passed in.
Graham McVicker's avatar
Graham McVicker committed
849
  Returntype : Bio::EnsEMBL::Slice
850 851 852 853
  Exceptions : throw if the feature does not have an attached slice
               throw if feature argument is not provided
  Caller     : fetch_by_gene_stable_id, fetch_by_transcript_stable_id,
               fetch_by_gene_id, fetch_by_transcript_id
Graham McVicker's avatar
Graham McVicker committed
854 855 856

=cut

857 858
sub fetch_by_Feature{
  my ($self, $feature, $size) = @_;
Graham McVicker's avatar
Graham McVicker committed
859

860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891
  $size ||= 0;

  if(!ref($feature) || !$feature->isa('Bio::EnsEMBL::Feature')) {
    throw('Feature argument expected.');
  }

  my $slice = $feature->slice();
  if(!$slice || !$slice->isa('Bio::EnsEMBL::Slice')) {
    throw('Feature must be attached to a valid slice.');
  }

  my $fstart = $feature->start();
  my $fend   = $feature->end();
  if(!defined($fstart) || !defined($fend)) {
    throw('Feature must have defined start and end.');
  }

  #convert the feature slice coordinates to seq_region coordinates
  my $slice_start  = $slice->start();
  my $slice_end    = $slice->end();
  my $slice_strand = $slice->strand();
  if($slice_start != 1 || $slice_strand != 1) {
    if($slice_strand == 1) {
      $fstart = $fstart + $slice_start - 1;
      $fend   = $fend   + $slice_start - 1;
    } else {
      my $tmp_start = $fstart;
      $fstart = $slice_end - $fend      + 1;
      $fend   = $slice_end - $tmp_start + 1;
    }
  }

892 893 894 895
  ## Size may be stored as a %age of the length of the feature
  ## Size = 100% gives no context
  ## Size = 200% gives context - 50% the size of the feature either side of 
  ## feature
Web Admin's avatar
Web Admin committed
896 897 898

  $size = int( ($1-100)/200 * ($fend-$fstart+1) ) if( $size =~/([\d+\.]+)%/ );

899
  #return a new slice covering the region of the feature
900 901 902 903 904 905 906 907
  return Bio::EnsEMBL::Slice->new
    (-seq_region_name   => $slice->seq_region_name,
     -seq_region_length => $slice->seq_region_length,
     -coord_system      => $slice->coord_system,
     -start             => $fstart - $size,
     -end               => $fend + $size,
     -strand            => 1,
     -adaptor           => $self);
908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939
}



=head2 fetch_by_misc_feature_attribute

  Arg [1]    : string $attribute_type
               The code of the attribute type
  Arg [2]    : (optional) string $attribute_value
               The value of the attribute to fetch by
  Arg [3]    : (optional) int $size
               The amount of flanking region around the misc feature desired.
  Example    : $slice = $sa->fetch_by_misc_feature_attribute('superctg',
                                                             'NT_030871');
               $slice = $sa->fetch_by_misc_feature_attribute('synonym',
                                                             'AL00012311',
                                                             $flanking);
  Description: Fetches a slice around a MiscFeature with a particular
               attribute type and value. If no value is specified then
               the feature with the particular attribute is used.
               If no size is specified then 0 is used.
  Returntype : Bio::EnsEMBL::Slice
  Exceptions : Throw if no feature with the specified attribute type and value
               exists in the database
               Warning if multiple features with the specified attribute type
               and value exist in the database.
  Caller     : webcode

=cut

sub fetch_by_misc_feature_attribute {
  my ($self, $attrib_type_code, $attrib_value, $size) = @_;
Graham McVicker's avatar
Graham McVicker committed
940

941
  my $mfa = $self->db()->get_MiscFeatureAdaptor();
Web Admin's avatar
Web Admin committed
942

943 944
  my $feats = $mfa->fetch_all_by_attribute_type_value($attrib_type_code,
                                                   $attrib_value);
Web Admin's avatar
Web Admin committed
945

946 947 948 949
  if(@$feats == 0) {
    throw("MiscFeature with $attrib_type_code=$attrib_value does " .
          "not exist in DB.");
  }
Graham McVicker's avatar
Graham McVicker committed
950

951 952 953 954
  if(@$feats > 1) {
    warning("MiscFeature with $attrib_type_code=$attrib_value is " .
            "ambiguous - using first one found.");
  }
955

956
  my ($feat) = @$feats;
957

958
  return $self->fetch_by_Feature($feat, $size);
Graham McVicker's avatar
Graham McVicker committed
959 960
}

961
=head2 fetch_normalized_slice_projection
Graham McVicker's avatar
Graham McVicker committed
962

963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019
  Arg [1]    : Bio::EnsEMBL::Slice $slice
  Example    :  ( optional )
  Description: gives back a project style result. The returned slices 
               represent the areas to which there are symlinks for the 
               given slice. start, end show which area on given slice is 
               symlinked
  Returntype : [[start,end,$slice][]]
  Exceptions : none
  Caller     : BaseFeatureAdaptor

=cut


sub fetch_normalized_slice_projection {
  my $self = shift;
  my $slice = shift;

  my $slice_seq_region_id = $self->get_seq_region_id( $slice );

  my $result = $self->{'asm_exc_cache'}->{$slice_seq_region_id};

  if(!defined($result)) {
    my $sql = "
      SELECT seq_region_id, seq_region_start, seq_region_end,
             exc_type, exc_seq_region_id, exc_seq_region_start,
             exc_seq_region_end
        FROM assembly_exception
       WHERE seq_region_id = ?";

    my $sth = $self->prepare( $sql );
    $sth->execute( $slice_seq_region_id );
    $result = $sth->fetchall_arrayref();
    $self->{'asm_exc_cache'}->{$slice_seq_region_id} = $result;
  }

  my (@haps, @pars);

  foreach my $row (@$result) {
    my ( $seq_region_id, $seq_region_start, $seq_region_end,
         $exc_type, $exc_seq_region_id, $exc_seq_region_start,
         $exc_seq_region_end ) = @$row;

    # need overlapping PAR and all HAPs if any
    if( $exc_type eq "PAR" ) {
      if( $seq_region_start <= $slice->end() && 
          $seq_region_end >= $slice->start() ) {
        push( @pars, [ $seq_region_start, $seq_region_end, $exc_seq_region_id,
                       $exc_seq_region_start, $exc_seq_region_end ] );
      }
    } else {
      push( @haps, [ $seq_region_start, $seq_region_end, $exc_seq_region_id,
                     $exc_seq_region_start, $exc_seq_region_end ] );
    }
  }

  if(!@pars && !@haps) {
    #just return this slice, there were no haps or pars
1020
    return  [bless ( [1,$slice->length, $slice], "Bio::EnsEMBL::ProjectionSegment")];
1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061
  }

  my @syms;
  if( @haps > 1 ) {
    my @sort_haps = sort { $a->[1] <=> $b->[1] } @haps;
    throw( "More than one HAP region not supported yet" );
  } elsif( @haps == 1 ) {
    my $hap = $haps[0];

    my $seq_reg_slice = $self->fetch_by_seq_region_id($slice_seq_region_id);
    my $exc_slice = $self->fetch_by_seq_region_id( $hap->[2] );

    #
    # lengths of haplotype and reference in db may be different
    # we want to use the maximum possible length for the mapping
    # between the two systems
    #
    my $len1 = $seq_reg_slice->length();
    my $len2 = $exc_slice->length();
    my $max_len = ($len1 > $len2) ? $len1 : $len2;

    #the inserted region can differ in length, but mapped sections
    #need to be same lengths
    my $diff = $hap->[4] - $hap->[1];

    # we want the region of the haplotype INVERTED
    push( @syms, [ 1, $hap->[0]-1, $hap->[2], 1, $hap->[3] - 1 ] );
    push( @syms, [ $hap->[1]+1, $max_len - $diff, 
                   $hap->[2], $hap->[4] + 1, $max_len ] );   
  }

  # for now haps and pars should not be both there, but in theory we 
  # could handle it here by cleverly merging the pars into the existing syms,
  # for now just:
  push( @syms, @pars );

  my $mapper = Bio::EnsEMBL::Mapper->new( "sym", "org" );
  for my $sym ( @syms ) {
    $mapper->add_map_coordinates( $slice_seq_region_id, $sym->[0], $sym->[1],
                                  1, $sym->[2], $sym->[3], $sym->[4] );
  }
1062

1063 1064 1065
  my @linked = $mapper->map_coordinates( $slice_seq_region_id,
                                         $slice->start(), $slice->end(),
                                         $slice->strand(), "sym" );
1066

1067 1068 1069 1070 1071 1072
  # gaps are regions where there is no mapping to another region
  my $rel_start = 1;

  #if there was only one coord and it is a gap, we know it is just the
  #same slice with no overlapping symlinks
  if(@linked == 1 && $linked[0]->isa('Bio::EnsEMBL::Mapper::Gap')) {
1073
    return [bless( [1,$slice->length, $slice], "Bio::EnsEMBL::ProjectionSegment" )];
1074 1075 1076 1077 1078 1079
  }

  my @out;
  for my $coord ( @linked ) {
    if( $coord->isa( "Bio::EnsEMBL::Mapper::Gap" )) {
      my $exc_slice = Bio::EnsEMBL::Slice->new
1080 1081 1082 1083 1084 1085 1086
        (-START             => $coord->start(),
         -END               => $coord->end(),
         -STRAND            => $slice->strand(),
         -COORD_SYSTEM      => $slice->coord_system(),
         -ADAPTOR           => $self,
         -SEQ_REGION_NAME   => $slice->seq_region_name(),
         -SEQ_REGION_LENGTH => $slice->seq_region_length());
1087 1088
      push( @out, bless ( [ $rel_start, $coord->length()+$rel_start-1,
                        $exc_slice ], "Bio::EnsEMBL::ProjectionSegment") );
1089 1090 1091 1092
    } else {
      my $exc_slice = $self->fetch_by_seq_region_id( $coord->id() );
      my $exc2_slice = Bio::EnsEMBL::Slice->new
        (
1093 1094 1095 1096 1097 1098 1099
         -START             => $coord->start(),
         -END               => $coord->end(),
         -STRAND            => $coord->strand(),
         -SEQ_REGION_NAME   => $exc_slice->seq_region_name(),
         -SEQ_REGION_LENGTH => $exc_slice->seq_region_length(),
         -COORD_SYSTEM      => $exc_slice->coord_system(),
         -ADAPTOR           => $self
1100 1101
        );
	
1102 1103
      push( @out, bless( [ $rel_start, $coord->length() + $rel_start - 1,
                    $exc2_slice ], "Bio::EnsEMBL::ProjectionSegment") );
1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125
    }
    $rel_start += $coord->length();
  }

  return \@out;
}




=head2 store

  Arg [1]    : Bio::EnsEMBL::Slice $slice
  Arg [2]    : (optional) $seqref reference to a string
               The sequence associated with the slice to be stored.
  Example    : $slice = Bio::EnsEMBL::Slice->new(...);
               $seq_region_id = $slice_adaptor->store($slice, \$sequence);
  Description: This stores a slice as a sequence region in the database
               and returns the seq region id. The passed in slice must
               start at 1, and must have a valid seq_region name and coordinate
               system. The attached coordinate system must already be stored in
               the database.  The sequence region is assumed to start at 1 and
1126 1127
               to have a length equalling the length of the slice.  The end of
               the slice must equal the seq_region_length.
1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143
               If the slice coordinate system is the sequence level coordinate
               system then the seqref argument must also be passed.  If the
               slice coordinate system is NOT a sequence level coordinate
               system then the sequence argument cannot be passed.
  Returntype : int 
  Exceptions : throw if slice has no coord system.
               throw if slice coord system is not already stored.
               throw if slice coord system is seqlevel and no sequence is 
                     provided.
               throw if slice coord system is not seqlevel and sequence is
                     provided.
               throw if slice does not start at 1
               throw if sequence is provided and the sequence length does not
                     match the slice length.
               throw if the SQL insert fails (e.g. on duplicate seq region)
               throw if slice argument is not passed
1144
               throw if the slice end is not equal to seq_region_length
1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175
  Caller     : database loading scripts

=cut



sub store {
  my $self = shift;
  my $slice = shift;
  my $seqref = shift;

  #
  # Get all of the sanity checks out of the way before storing anything
  #

  if(!ref($slice) || !$slice->isa('Bio::EnsEMBL::Slice')) {
    throw('Slice argument is required');
  }

  my $cs = $slice->coord_system();
  throw("Slice must have attached CoordSystem.") if(!$cs);

  my $db = $self->db();
  if(!$cs->is_stored($db)) {
    throw("Slice CoordSystem must already be stored in DB.") 
  }

  if($slice->start != 1 || $slice->strand != 1) {
    throw("Slice must have start==1 and strand==1.");
  }

1176 1177 1178 1179
  if($slice->end() != $slice->seq_region_length()) {
    throw("Slice must have end==seq_region_length");
  }

1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227