Make exon ID unique by gene, not by transcript

miniannotator.py

@@ -301,6 +301,72 @@ class Miniannotator:
 
         return full_genes
 
+    @staticmethod
+    def _rename_exons(gene):
+        """
+        Rename exons: exons of a same gene which have the same coordinates must have the same name
+        special name for two exons which overlap but does not have the same coordinates
+
+        :param gene: First item: gene id. Second item: gene object, containing transcripts, exons and indels
+        :type gene: list
+        :return: gene with exons renamed
+        :rtype: str, dict
+        """
+        gene_id = gene[0]
+        gene_o = gene[1]
+        exons_corr = {}
+        exons_by_coord = {}
+
+        # Save exons by coordinates
+        for transcript in gene_o["transcripts"].values():
+            for ex_id, exon in transcript["exons"].items():
+                coords = (exon["start"], exon["end"])
+                if coords in exons_by_coord:
+                    exons_by_coord[coords].append(ex_id)
+                else:
+                    exons_by_coord[coords] = [ex_id]
+
+        # Build new exons name
+        ex_order = sorted(exons_by_coord.keys())
+        i = 0
+        ex_nb = 1
+        for coord in ex_order:
+            create_new = True
+            if i > -1:
+                for k in range(i, 0, -1):
+                    o_coord = ex_order[k]
+                    if coord[0] < o_coord[0] < coord[1] or coord[0] < o_coord[1] < coord[1] or \
+                            o_coord[0] < coord[0] < o_coord[1] or o_coord[0] < coord[1] < o_coord[1]:
+                        n_alt = 2
+                        new_id_basis = exons_corr[exons_by_coord[o_coord][0]].split("_")[0]
+                        new_id = new_id_basis + "_%d" % n_alt
+                        while new_id in exons_corr.values():
+                            n_alt += 1
+                            new_id = new_id_basis + "_%d" % n_alt
+                        create_new = False
+                        break
+            if create_new:
+                new_id = gene_id + ".e%d" % ex_nb
+                ex_nb += 1
+            for ex_id in exons_by_coord[coord]:
+                exons_corr[ex_id] = new_id
+
+            i += 1
+
+        # Rename exons in transcripts
+        for transcript in gene_o["transcripts"].values():
+            # Rename exons:
+            new_exons = OrderedDict()
+            for ex_id, exon in transcript["exons"].items():
+                new_exons[exons_corr[ex_id]] = exon
+            transcript["exons"] = new_exons
+
+            # Rename exons in indels:
+            for indel in transcript["indels"]:
+                indel["exon"] = exons_corr[indel["exon"]]
+
+        return gene_id, gene_o
+
     def write_gtf_file(self, gtf_file, genes):
         """
         Write genes, transcripts, exons and indels to a GTF file
@@ -400,8 +466,12 @@ class Miniannotator:
 
         full_genes = self._build_genes_transcripts(genes=genes, exons=exons, indels=indels)
 
+        print("Renaming exons...", flush=True)
+        pool = multiprocessing.Pool(processes=self.conf["threads"])
+        full_genes_2 = OrderedDict(pool.map(Miniannotator._rename_exons, list(full_genes.items())))
+
         print("Writing to GTF file...", flush=True)
-        self.write_gtf_file(gtf_file=gtf_file, genes=full_genes)
+        self.write_gtf_file(gtf_file=gtf_file, genes=full_genes_2)
 
 
 if __name__ == "__main__":