Fix unstable variants in models output

CHANGELOG.md

@@ -8,6 +8,10 @@ Please add a new candidate release at the top after changing the latest one. Fee
 
 Try to use the following format:
 
+## [3.8.3]
+
+- Fixed unstable compounds order in models output ([#133](https://github.com/Clinical-Genomics/genmod/issues/133))
+
 ## [3.8.0]
 - Rank score normalisation
 
@@ -22,4 +26,4 @@ Try to use the following format:
 - Adds support for escaped characters in FORMAT description header strings
 
 ### Fixed
-- Documentation about cli options `strict` and `phased`
+- Documentation about cli options `strict` and `phased`

genmod/score_variants/compound_scorer.py

@@ -267,6 +267,8 @@ def run(self):
                             new_compound = "{0}>{1}".format(compound_id, compound_score)
                             scored_compound_list.append(new_compound)
 
+                        # Sort compound variants lexicographically
+                        scored_compound_list.sort()
                         new_compound_string = "{0}:{1}".format(
                             compound_family_id, '|'.join(scored_compound_list))
 

requirements.txt

@@ -7,3 +7,4 @@ configobj == 5.0.8
 intervaltree == 3.1.0
 extract_vcf == 0.5
 vcftoolbox == 1.5.1
+six == 1.16.0

setup.py

@@ -36,7 +36,8 @@
         'configobj >= 5.0.8',
         'intervaltree >= 3.1.0',
         'extract_vcf >= 0.5',
-        'vcftoolbox >= 1.5.1'
+        'vcftoolbox >= 1.5.1',
+        'six >= 1.16.0',
     ],
     packages=find_packages(
         exclude=('tests*', 'docs', 'examples', 'configs')