testing stable id mapping code changes where biotype changes are penalised...

testing stable id mapping code changes where biotype changes are penalised when first calculating the transcript score

modules/Bio/EnsEMBL/IdMapping/Cache.pm

@@ -1037,6 +1037,18 @@ sub get_DBAdaptor {
 } ## end sub get_DBAdaptor
 
 
+sub get_production_DBAdaptor() {
+  my ($self) = @_;
+  my $dba = new Bio::EnsEMBL::DBSQL::DBAdaptor(
+                       -host   => $self->conf->param("productionhost"),
+                       -port   => $self->conf->param("productionport"),
+                       -user   => $self->conf->param("productionuser"),
+                       -pass   => $self->conf->param("productionpass"),
+                       -dbname => $self->conf->param("productiondbname"));
+  return $dba;
+}
+
+
 sub cache_file_exists {
   my $self = shift;
   my $type = shift;

modules/Bio/EnsEMBL/IdMapping/InternalIdMapper/BaseMapper.pm

@@ -90,18 +90,18 @@ sub basic_mapping {
     sort { $b->score <=> $a->score } @{ $matrix->get_all_Entries };
 
   # debug
-  #my $idx = substr($mapping_name, -1);
+  my $idx = substr($mapping_name, -1);
 
   while ( my $entry = shift(@sorted_entries) ) {
 
-    #$self->logger->debug("\nxxx$idx ".$entry->to_string." ");
+    $self->logger->debug("\nxxx$idx ".$entry->to_string." ");
 
     # we already found a mapping for either source or target
     next
       if (    $sources_done->{ $entry->source }
            or $targets_done->{ $entry->target } );
 
-    #$self->logger->debug('d');
+    $self->logger->debug('d');
 
     # there's a better mapping for either source or target
     next
@@ -109,16 +109,17 @@ sub basic_mapping {
                            $entry, $matrix, $sources_done, $targets_done
            ) );
 
-    #$self->logger->debug('h');
+    $self->logger->debug('h');
 
     # check for ambiguous mappings; they are dealt with later
+    # unless all the other ambiguous mappings have already been dealt with
     my $other_sources = [];
     my $other_targets = [];
 
     if ( $self->ambiguous_mapping( $entry,         $matrix,
                                    $other_sources, $other_targets ) )
     {
-      #$self->logger->debug('a');
+      $self->logger->debug('a');
 
       $other_sources =
         $self->filter_sources( $other_sources, $sources_done );
@@ -128,7 +129,7 @@ sub basic_mapping {
       next if ( scalar(@$other_sources) or scalar(@$other_targets) );
     }
 
-    #$self->logger->debug('A');
+    $self->logger->debug('A');
 
     # this is the best mapping, add it
     $mappings->add_Entry($entry);

modules/Bio/EnsEMBL/IdMapping/TranscriptScoreBuilder.pm

@@ -317,20 +317,36 @@ sub score_matrix_from_flag_matrix {
 
         } else {
 
-=cut
+          my $source_transcript_biotype_group = $self->get_biotype_group($source_transcript->biotype());
+          my $target_transcript_biotype_group = $self->get_biotype_group($target_transcript->biotype());
+
           # debug
           $self->logger->info($source_transcript->id.":".$target_transcript->id.
             " source score: $source_transcript_score".
             " source length: $source_transcript_length".
+            " source biotype:" . $source_transcript->biotype() .
+            " source group: $source_transcript_biotype_group".
             " target score: $target_transcript_score".
+            " target biotype:" . $target_transcript->biotype() .
+            " target group: $target_transcript_biotype_group".
             " target length: $target_transcript_length\n");
-=cut
 
-          # everything is fine, add score to matrix
+          
           my $transcript_score =
             ($source_transcript_score + $target_transcript_score) /
             ($source_transcript_length + $target_transcript_length);
 
+## Add penalty if biotypes are different
+          if ($source_transcript->biotype() ne $target_transcript->biotype()) {
+            $transcript_score = $transcript_score * 0.9;
+          }
+
+## Add penalty if biotype groups are different
+          if ($source_transcript_biotype_group ne $target_transcript_biotype_group) {
+            $transcript_score = $transcript_score * 0.8;
+          }
+
+          # everything is fine, add score to matrix
           if ($transcript_score > $transcript_score_threshold) {
             $matrix->add_score($source_transcript->id, $target_transcript->id,
               $transcript_score);
@@ -494,5 +510,22 @@ sub non_mapped_gene_rescore {
 } ## end sub non_mapped_gene_rescore
 
 
+sub get_biotype_group {
+  my ($self, $biotype) = @_;
+  my $dba = $self->cache->get_production_DBAdaptor();
+  my $helper   = $self->cache->get_production_DBAdaptor()->dbc()->sql_helper();
+
+  my $sql      = q{
+     SELECT biotype_group
+     FROM biotype
+     WHERE object_type = 'transcript'
+     AND is_current = 1
+     AND name = ?
+     AND db_type like '%core%' };
+  my $result = $helper->execute_simple(-SQL => $sql, -PARAMS => [$biotype]);
+  return $result->[0];
+}
+
+  
 1;