Updated moalmanac db to 2024-04-11 release (#19)

moalmanac/config.ini

@@ -9,7 +9,7 @@ plot_preclinical_efficacy = on
 
 [versions]
 interpreter = 0.6.0
-database = v.2023-11-09
+database = v.2024-04-11
 
 [exac]
 exac_common_af_threshold = 0.001

moalmanac/datasources/moalmanac/README.md

@@ -7,12 +7,14 @@ The Molecular Oncology Almanac attempts to capture the current body of knowledge
 Several other services exist within the Molecular Oncology Almanac ecosystem. See [this repository's docs folder](/docs/) for more information.
 
 ## Usage: Formatting the database for use
-This method uses a document-based format of the database, which is built using the [database repository](https://github.com/vanallenlab/moalmanac-db) and `create_almanac_db.py`. If MOAlmanac is updated, **please also regenerate [preclinical datasources](../preclinical/)**.
+This method uses a json-based format of the database, which is built using the [database repository](https://github.com/vanallenlab/moalmanac-db) and `create_almanac_db.py`. If MOAlmanac is updated, **please also regenerate [preclinical datasources](../preclinical/)**.
 
 Arguments:
 ```
-    --directory, -d     <string>    path to the moalmanac-db repository's content folder
-    --version, -v       <string>    release name for the database content being used, should match the release from moalmanac-db
+    --config, -c        <string>    path to config.ini file, "../../config.ini" by default
+    --file, -f          <string>    path to moalmanac-db json file
+    --release, -r       <string>    date of moalmanac-db release, should match the release from moalmanac-db
+    --version, -v       <string>    database version of the moalmanac-db schema
 ```
 
 This should be run with this repository's virtual environment enabled. 

moalmanac/datasources/moalmanac/create_almanac_db.py

@@ -3,25 +3,6 @@
 import json
 import sys
 
-
-def create_config():
-    config = configparser.ConfigParser()
-    config.read('../../config.ini')
-    return config
-
-
-CONFIG = create_config()
-feature_type_section = 'feature_types'
-FEATURE_TYPE_SOMATIC = CONFIG[feature_type_section]['mut']
-FEATURE_TYPE_GERMLINE = CONFIG[feature_type_section]['germline']
-FEATURE_TYPE_COPY_NUMBER = CONFIG[feature_type_section]['cna']
-FEATURE_TYPE_FUSION = CONFIG[feature_type_section]['fusion']
-FEATURE_TYPE_MUTATIONAL_BURDEN = CONFIG[feature_type_section]['burden']
-FEATURE_TYPE_MUTATIONAL_SIGNATURE = CONFIG[feature_type_section]['signature']
-FEATURE_TYPE_MICROSATELLITE_STABILITY = CONFIG[feature_type_section]['microsatellite']
-FEATURE_TYPE_ANEUPLOIDY = CONFIG[feature_type_section]['aneuploidy']
-FEATURE_TYPE_KNOCKDOWN = CONFIG[feature_type_section]['knockdown']
-
 ASSERTION_FIELDS = [
     'disease', 'context', 'oncotree_term', 'oncotree_code',
     'therapy_name', 'therapy_strategy', 'therapy_type',
@@ -135,42 +116,59 @@ def somatic_variant(cls, record):
             return f"{gene}"
 
     @classmethod
-    def generate(cls, record):
+    def generate(cls, record, config_dictionary):
+        feature_type_dictionary = generate_feature_type_string_lookup_dictionary(config_dictionary)
         feature_type = record['feature_type']
-        if feature_type == FEATURE_TYPE_SOMATIC:
+        if feature_type == feature_type_dictionary['somatic_variants']:
             return cls.somatic_variant(record)
-        elif feature_type == FEATURE_TYPE_GERMLINE:
+        elif feature_type == feature_type_dictionary['germline_variants']:
             return cls.germline_variant(record)
-        elif feature_type == FEATURE_TYPE_COPY_NUMBER:
+        elif feature_type == feature_type_dictionary['copy_number_alterations']:
             return cls.copy_number(record)
-        elif feature_type == FEATURE_TYPE_FUSION:
+        elif feature_type == feature_type_dictionary['fusions']:
             return cls.rearrangements(record)
-        elif feature_type == FEATURE_TYPE_ANEUPLOIDY:
+        elif feature_type == feature_type_dictionary['aneuploidy']:
             return cls.aneuploidy(record)
-        elif feature_type == FEATURE_TYPE_MUTATIONAL_BURDEN:
+        elif feature_type == feature_type_dictionary['tmb']:
             return cls.mutational_burden(record)
-        elif feature_type == FEATURE_TYPE_MICROSATELLITE_STABILITY:
+        elif feature_type == feature_type_dictionary['microsatellite']:
             return cls.microsatellite_stability(record)
-        elif feature_type == FEATURE_TYPE_MUTATIONAL_SIGNATURE:
+        elif feature_type == feature_type_dictionary['signature']:
             return cls.cosmic_mutational_signature(record)
-        elif feature_type == FEATURE_TYPE_KNOCKDOWN:
+        elif feature_type == feature_type_dictionary['knockdown']:
             return cls.knockdown(record)
         else:
             print(f'ERROR: {feature_type} does not have a feature display format function.')
 
 
+def generate_feature_type_string_lookup_dictionary(config):
+    feature_type_section = 'feature_types'
+    dictionary = {
+        'somatic_variants': config[feature_type_section]['mut'],
+        'germline_variants': config[feature_type_section]['germline'],
+        'copy_number_alterations': config[feature_type_section]['cna'],
+        'fusions': config[feature_type_section]['fusion'],
+        'tmb': config[feature_type_section]['burden'],
+        'signature': config[feature_type_section]['signature'],
+        'microsatellite': config[feature_type_section]['microsatellite'],
+        'aneuploidy': config[feature_type_section]['aneuploidy'],
+        'knockdown': config[feature_type_section]['knockdown'],
+    }
+    return dictionary
+
+
 def check_fields(fields, record):
     for field in fields:
         if field not in record.keys():
             sys.exit(f'{field} not present in record.\n{record}')
 
 
-def check_format(record):
+def check_format(record, config_dictionary):
     check_fields(SOURCE_FIELDS, record)
     check_fields(ASSERTION_FIELDS, record)
 
     feature_type = record['feature_type']
-    feature_type_fields = get_feature_type_fields(feature_type)
+    feature_type_fields = get_feature_type_fields(feature_type, config_dictionary)
     check_fields(feature_type_fields, record)
 
 
@@ -186,24 +184,25 @@ def extract_genes(records):
     return genes_sorted
 
 
-def get_feature_type_fields(feature_type):
-    if feature_type == FEATURE_TYPE_SOMATIC:
+def get_feature_type_fields(feature_type, config_dictionary):
+    feature_type_dictionary = generate_feature_type_string_lookup_dictionary(config_dictionary)
+    if feature_type == feature_type_dictionary['somatic_variants']:
         return VARIANT_FIELDS
-    elif feature_type == FEATURE_TYPE_GERMLINE:
+    elif feature_type == feature_type_dictionary['germline_variants']:
         return GERMLINE_FIELDS
-    elif feature_type == FEATURE_TYPE_COPY_NUMBER:
+    elif feature_type == feature_type_dictionary['copy_number_alterations']:
         return COPY_NUMBER_FIELDS
-    elif feature_type == FEATURE_TYPE_FUSION:
+    elif feature_type == feature_type_dictionary['fusions']:
         return REARRANGEMENT_FIELDS
-    elif feature_type == FEATURE_TYPE_MUTATIONAL_BURDEN:
+    elif feature_type == feature_type_dictionary['tmb']:
         return MUTATIONAL_BURDEN_FIELDS
-    elif feature_type == FEATURE_TYPE_MUTATIONAL_SIGNATURE:
+    elif feature_type == feature_type_dictionary['signature']:
         return MUTATIONAL_SIGNATURE_FIELDS
-    elif feature_type == FEATURE_TYPE_MICROSATELLITE_STABILITY:
+    elif feature_type == feature_type_dictionary['microsatellite']:
         return MICROSATELLITE_FIELDS
-    elif feature_type == FEATURE_TYPE_ANEUPLOIDY:
+    elif feature_type == feature_type_dictionary['aneuploidy']:
         return ANEUPLOIDY_FIELDS
-    elif feature_type == FEATURE_TYPE_KNOCKDOWN:
+    elif feature_type == feature_type_dictionary['knockdown']:
         return KNOCKDOWN_FIELDS
     else:
         sys.exit(f'feature type {feature_type} present in database but not accounted for.')
@@ -218,6 +217,12 @@ def initialize():
     }
 
 
+def load_config(file_path):
+    config = configparser.ConfigParser()
+    config.read(file_path)
+    return config
+
+
 def load_json(json_file):
     with open(json_file, 'r') as f:
         json_data = json.load(f)
@@ -229,7 +234,7 @@ def write_json(file, data):
         json.dump(data, f, ensure_ascii=False, indent=4)
 
 
-def main(version, release, content):
+def main(version, release, content, config):
     db = initialize()
     db['release'] = release
     db['version'] = version
@@ -238,23 +243,48 @@ def main(version, release, content):
     db['genes'] = db_genes
 
     for record in content:
-        check_format(record)
-        record['feature_display'] = DisplayAlteration.generate(record)
+        if record['_deprecated']:
+            continue
+        check_format(record, config)
+        record['feature_display'] = DisplayAlteration.generate(record, config)
     db['content'] = content
     write_json('molecular-oncology-almanac.json', db)
 
 
 if __name__ == "__main__":
-    arg_parser = argparse.ArgumentParser(prog='Create Molecular Oncology Almanac datasource',
-                                         description='Compiles MOAlmanac db for use with algorithm')
-    arg_parser.add_argument('--file', '-f',
-                            help='Molecular Oncology Almanac db file from https://github.com/vanallenlab/moalmanac-db')
-    arg_parser.add_argument('--version', '-v',
-                            help='Database version; e.g. 1.0.0')
-    arg_parser.add_argument('--release', '-r',
-                            help='Database content release; e.g. v.2022-12-01')
+    arg_parser = argparse.ArgumentParser(
+        prog='Create Molecular Oncology Almanac datasource',
+        description='Compiles MOAlmanac db for use with algorithm'
+    )
+    arg_parser.add_argument(
+        '--config',
+        '-c',
+        help='MOAlmanac configuration file',
+        default='../../config.ini'
+    )
+    arg_parser.add_argument(
+        '--file',
+        '-f',
+        help='Molecular Oncology Almanac db file from https://github.com/vanallenlab/moalmanac-db'
+    )
+    arg_parser.add_argument(
+        '--release',
+        '-r',
+        help='Database content release; e.g. v.2022-12-01'
+    )
+    arg_parser.add_argument(
+        '--version',
+        '-v',
+        help='Database version; e.g. 1.0.0'
+    )
     args = arg_parser.parse_args()
     print(args)
 
     moalmanac_json = load_json(args.file)
-    main(args.version, args.release, moalmanac_json)
+    config_ini = load_config(args.config)
+    main(
+        version=args.version,
+        release=args.release,
+        content=moalmanac_json,
+        config=config_ini
+    )