merge overlapping domain locations

run_domain2go_app.py

@@ -4,6 +4,7 @@ from Bio import SeqIO
 import os
 import time
 import pandas as pd
+import intervaltree
 
 def find_domains(email, sequence, name):
 
@@ -72,10 +73,10 @@ def find_domains(email, sequence, name):
                         entries[entry['accession']]['locations'].extend(location_list)
 
                 entries[entry['accession']]['locations'] = list(set(entries[entry['accession']]['locations']))
-                entries[entry['accession']]['locations'] = sorted([i.split('-') for i in entries[entry['accession']]['locations']], key=lambda x: (int(x[0]), int(x[1])))
-                entries[entry['accession']]['locations'] = ['-'.join(i) for i in entries[entry['accession']]['locations']]
-                # entries[entry['accession']]['locations'] = '|'.join(entries[entry['accession']]['locations'])
-        
+                if len(entries[entry['accession']]['locations']) > 1:
+                    entries[entry['accession']]['locations'] = merge_locations(entries[entry['accession']]['locations'])
+                    entries[entry['accession']]['locations'] = sorted([i.split('-') for i in entries[entry['accession']]['locations']], key=lambda x: (int(x[0]), int(x[1])))
+                    entries[entry['accession']]['locations'] = ['-'.join(i) for i in entries[entry['accession']]['locations']]
     if entries:
         result_text  = 'Domains found.'
 
@@ -92,6 +93,14 @@ def find_domains(email, sequence, name):
                         
     # generate protein function predictions based on domain2go mappings
 
+
+def merge_locations(locations):
+    temp_locs = [i.split('-') for i in locations]
+    tree = intervaltree.IntervalTree.from_tuples(temp_locs)
+    tree.merge_overlaps()
+    merged_locations = ['-'.join([i.begin, i.end]) for i in tree]
+    return merged_locations
+    
 def generate_function_predictions(domains_df, mapping_path):
     
     # read domain2go mappings
@@ -115,4 +124,4 @@ def generate_function_predictions(domains_df, mapping_path):
         # save protein function predictions
         protein_name = domains_df['protein_name'].iloc[0]
         result_text= 'Function predictions found.'
-        return [result_text, merged_df]
+        return [result_text, merged_df]