Update VEP JSON output for var_synonyms (#915)

Author: aparton

modules/Bio/EnsEMBL/VEP/OutputFactory/JSON.pm

@@ -121,6 +121,7 @@ my %NUMBERIFY_EXEMPT = (
 my @LIST_FIELDS = qw(
   clin_sig
   pubmed
+  var_synonyms
 );
 
 
@@ -418,7 +419,19 @@ sub add_colocated_variant_info_JSON {
 
   # lists
   foreach my $field(grep {defined($ex->{$_})} @LIST_FIELDS) {
-    $ex->{$field} = [split(',', $ex->{$field})];
+    if($field eq 'var_synonyms'){
+      my @source_list = split('--', $ex->{$field});
+      my %output_hash = ();
+      foreach my $syn_source(@source_list){
+        my @split = split('::', $syn_source);
+        my @syns = split(',', $split[1]);
+	$output_hash{$split[0]} = \@syns;
+      }
+      $ex->{$field} = \%output_hash;
+    }
+    else{
+      $ex->{$field} = [split(',', $ex->{$field})];
+    }
   }
 
   push @{$hash->{colocated_variants}}, $ex;

t/CacheDir.t

@@ -205,7 +205,8 @@ is_deeply(
       'gnomAD_FIN',
       'gnomAD_NFE',
       'gnomAD_OTH',
-      'gnomAD_SAS'
+      'gnomAD_SAS',
+      'var_synonyms'
     ],
     'assembly' => 'GRCh38',
     'valid_chromosomes' => [21, 22, 'LRG_485'],

t/OutputFactory_JSON.t

@@ -156,6 +156,9 @@ SKIP: {
           'start' => 25585733,
           'strand' => 1,
           'allele_string' => 'C/T',
+	  'var_synonyms' => {
+	    'ClinVar' => ['TEST']
+	  },
           'pubmed' => [10, 20, 30],
           'clin_sig' => ["pathogenic", "benign"],
         }
@@ -288,6 +291,9 @@ SKIP: {
           'start' => 25585733,
           'strand' => 1,
           'allele_string' => 'C/T',
+          'var_synonyms' => {
+            'ClinVar' => ['TEST']
+          },
         }
       ],
       'assembly_name' => 'GRCh38',
@@ -382,7 +388,7 @@ SKIP: {
       'id' => 'rs142513484',
       'allele_string' => 'C/T',
       'most_severe_consequence' => 'missense_variant',
-      'start' => 25585733
+      'start' => 25585733,
     },
     'get_all_lines_by_InputBuffer - everything'
   );

t/testdata/cache/homo_sapiens/84_GRCh38/21/25000001-26000000_var.gz

@@ -12,7 +12,7 @@ source_gencode	GENCODE 24
 source_genebuild	2014-07
 source_regbuild	13.0
 source_assembly	GRCh38.p5
-variation_cols	variation_name,failed,somatic,start,end,allele_string,strand,minor_allele,minor_allele_freq,clin_sig,phenotype_or_disease,pubmed,AFR,AMR,EAS,EUR,SAS,AA,EA,gnomAD,gnomAD_AFR,gnomAD_AMR,gnomAD_ASJ,gnomAD_EAS,gnomAD_FIN,gnomAD_NFE,gnomAD_OTH,gnomAD_SAS
+variation_cols	variation_name,failed,somatic,start,end,allele_string,strand,minor_allele,minor_allele_freq,clin_sig,phenotype_or_disease,pubmed,AFR,AMR,EAS,EUR,SAS,AA,EA,gnomAD,gnomAD_AFR,gnomAD_AMR,gnomAD_ASJ,gnomAD_EAS,gnomAD_FIN,gnomAD_NFE,gnomAD_OTH,gnomAD_SAS,var_synonyms
 source_COSMIC	80
 source_HGMD-PUBLIC	20164
 source_ESP	20141103