Merge pull request #380 from eweitz/visualize-variant-cache

Show clinical variants on gene models

Author: eweitz

dist/data/cache/gene-structures/homo-sapiens-gene-structures.tsv.gz

@@ -16,10 +16,10 @@ module.exports = function(config) {
     // list of files / patterns to load in the browser
     files: [
       'src/js/index.js',
-      'test/offline/**.test.js',
-      'test/online/**.test.js',
+      // 'test/offline/**.test.js',
+      // 'test/online/**.test.js',
       // 'test/online/related-genes.test.js',
-      // 'test/offline/gene-structure.test.js',
+      'test/offline/gene-structure.test.js',
       // 'test/offline/tissue.test.js',
       {pattern: 'dist/data/**', watched: false, included: false, served: true, nocache: false}
     ],

dist/data/cache/variants/homo-sapiens-variants.tsv.gz

@@ -262,6 +262,8 @@ def parse_mrna(raw_mrna, biotypes_list):
 def build_structures(structures_by_id):
     biotypes_list = list(biotypes.keys())
 
+    prev_gene = ''
+
     structures = []
     for id in structures_by_id:
         structure_lists = structures_by_id[id]
@@ -273,8 +275,19 @@ def build_structures(structures_by_id):
 
         for structure_list in structure_lists[1:]:
             subpart = parse_transcript_subpart(structure_list, mrna_start)
-            structure += [";".join(subpart) ]
-
+            structure += [";".join(subpart)]
+
+        # Set transcript start coordinate relative to most-upstream transcript
+        # This enables projecting genomic features (e.g. variants) onto
+        # transcript coordinates.  It also enables viewing multiple transcripts
+        # in genomic coordinates, like typical genome browsers (Ensembl, IGV).
+        gene_name = structure[1].split('-')[0]
+        if gene_name != prev_gene:
+            gene_start = int(mrna_start) # Start of 1st transcript is gene start
+            prev_gene = gene_name
+        mrna_start_offset = str(int(mrna_start) - gene_start)
+
+        structure.insert(2, mrna_start_offset)
         structures.append(structure)
 
     return structures
@@ -288,8 +301,8 @@ def parse_structures(canonical_ids, gff_path, gff_url):
 
     Parts of a transcript that comprise "gene structure" here:
         * Exons: regions of gene not removed by RNA splicing
-        * 3'-UTR: Three prime untranslated region; start region
-        * 5'-UTR: Fix prime untranslated region; end region
+        * 5'-UTR: Fix prime untranslated region; start region (for +, end for -)
+        * 3'-UTR: Three prime untranslated region; end region (for +, start for -)
 
     (Introns are the regions between 3'- and 5'-UTRs that are not exons.
     These are implied in the structure, and not modeled explicitly.)