Merge pull request #384 from dglemos/hgvs/fix_refseq_vep_96

Fix HGVS refseq transcripts
Ensembl · May 22, 2019 · 617872b · 617872b
2 parents 1f67fdd + fbf2183
commit 617872b
Show file tree

Hide file tree

Showing 2 changed files with 24 additions and 2 deletions.
diff --git a/modules/Bio/EnsEMBL/Variation/DBSQL/VariationFeatureAdaptor.pm b/modules/Bio/EnsEMBL/Variation/DBSQL/VariationFeatureAdaptor.pm
@@ -1784,6 +1784,10 @@ sub fetch_by_hgvs_notation {
     my $transcript = $transcript_adaptor->fetch_by_stable_id($reference);
 
     my @transcripts;
+
+    if(!defined($transcript) && ($reference =~ /NM_/ || $reference =~ /XM_/)){
+      @transcripts = @{$transcript_adaptor->fetch_all_by_external_name($reference)};
+    }
 
     #try and fetch via gene
     if(!defined($transcript)) {
@@ -1851,6 +1855,11 @@ sub fetch_by_hgvs_notation {
 
     my @transcripts;
 
+    # Fetch xref transcript
+    if(!defined($transcript) && ($reference =~ /NP/ || $reference =~ /XP/)){
+      @transcripts = @{$transcript_adaptor->fetch_all_by_external_name($reference)};
+    }
+
     # support some malformed HGVS
     if(!defined($transcript)) {
       # Seeing transcripts erroneously submitted with p. changes

diff --git a/modules/Bio/EnsEMBL/Variation/TranscriptVariationAllele.pm b/modules/Bio/EnsEMBL/Variation/TranscriptVariationAllele.pm
@@ -952,6 +952,7 @@ sub hgvs_transcript_reference{
 
 }
 
+use Data::Dumper;
 
 =head2 hgvs_protein
 
@@ -1022,8 +1023,20 @@ sub hgvs_protein {
 
   ## checks complete - start building term
 
-  ### get reference sequence and add seq version unless LRG
-  $hgvs_notation->{ref_name} = $tr->translation->display_id();
+  ### get reference sequence
+  ### this is a temporary fix
+  if($tr->stable_id =~ /^ENS|^LRG/ || !defined($tr->analysis) || (defined($tr->analysis) && !defined($tr->analysis->db()))){
+    $hgvs_notation->{ref_name} = $tr->translation->display_id();
+  }
+  else{
+    ### get RefSeq identifiers
+    my @entries = grep {$_->{dbname} eq 'GenBank'} @{$tr->translation->get_all_DBEntries};
+      if(scalar @entries == 1){
+        $hgvs_notation->{ref_name} = $entries[0]->{primary_id};
+      }
+   }
+
+  # Add seq version unless LRG
   $hgvs_notation->{ref_name} .= "." . $tr->translation->version() 
     unless ($hgvs_notation->{ref_name}=~ /\.\d+$/ || $hgvs_notation->{ref_name} =~ /LRG/);