Merge efe8229 into c5ab8db

scripts/importer/data_importer/raw_data_importer.py

@@ -199,30 +199,38 @@ def _parse_manta(self):
                     continue
 
                 if base["chrom"].startswith('GL') or base["chrom"].startswith('MT'):
-                    # TODO keep this?
+                    # A BND from GL or MT.
                     continue
 
                 if 'NSAMPLES' in info:
                     # save this unless we already know the sample size
                     samples = int(info['NSAMPLES'])
 
                 alt_alleles = base['alt'].split(",")
-                # TODO suspect for allelecount or callcount:
-                #    OCC,Number=1,Type=Integer,Description="The number of occurences of the event in the database"
                 for i, alt in enumerate(alt_alleles):
                     data = dict(base)
                     data['allele_freq'] = float(info.get('FRQ'))
                     data['alt'], data['mate_chrom'], data['mate_start'] = re.search('(.+)[[\]](.*?):(\d+)[[\]]', alt).groups()
                     if data['mate_chrom'].startswith('GL') or data['mate_chrom'].startswith('MT'):
+                        # A BND from a chromosome to GL or MT.
+                        # TODO ask a bioinformatician if these cases should be included or not
                         continue
-                    if 'MATEID' in info:
-                        data['mate_id'] = info['MATEID']
+                    data['mate_id'] = info.get('MATEID', '')
                     data['variant_id'] = '{}-{}-{}-{}'.format(data['chrom'], data['pos'], data['ref'], alt)
-
+                    data['allele_count'] = data.get('allele_count', 0)
+                    data['allele_num'] = data.get('allele_num', 0)
                     batch += [data]
-                    if self.settings.count_calls:
-                        self.get_callcount(data)  # count calls (one per reference)
-                        self.counter['beaconvariants'] += 1  # count variants (one per alternate)
+                    if self.settings.add_reversed_mates:
+                        # If the vcf only contains one line per breakend, add the reversed version to the database here.
+                        reversed = dict(data)
+                        # TODO Note: in general, ref and alt cannot be assumed to be the same in the reversed direction,
+                        # but our data (so far) only contains N, so we just keep them as is for now.
+                        reversed.update({'mate_chrom': data['chrom'], 'chrom': data['mate_chrom'],
+                                         'mate_start': data['pos'], 'pos': data['mate_start'],
+                                         'chrom_id': data['mate_id'], 'mate_id': data['chrom_id']})
+                        reversed['variant_id'] = '{}-{}-{}-{}'.format(reversed['chrom'], reversed['pos'], reversed['ref'], alt)
+                        # TODO should the `counter` be increased here?
+                        batch += [reversed]
 
                 counter += 1  # count variants (one per vcf row)
 
@@ -491,7 +499,7 @@ def start_import(self):
         if self.settings.add_mates:
             self._parse_manta()
             if self.settings.count_calls:
-                self._create_beacon_counts()
+                logging.warning('Do not know how to count calls in the manta file. Skipping this...')
         elif self.settings.variant_file:
             self._insert_variants()
             if self.settings.count_calls:

scripts/importer/importer.py

@@ -100,6 +100,8 @@
                               " the requirements"))
     PARSER.add_argument("--add_mates", action="store_true",
                         help=("Parse MANTA file and add the breakends to the db"))
+    PARSER.add_argument("--add_reversed_mates", action="store_true",
+                        help=("Assume input data only contain one line per BND, covering both directions"))
 
     ARGS = PARSER.parse_args()
 

sql/beacon_schema.sql

@@ -91,17 +91,17 @@ CREATE OR REPLACE VIEW beacon.beacon_mate_table AS
                           d.short_name,
                           av.dataset_version) AS datasetId,
            substr(dm.chrom, 1, 2) AS chromosome,
-           dm.pos - 1 AS "chromosomeStart",
+           dm.pos - 1 AS chromosomeStart,
            dm.chrom_id as chromosomePos,
            dm.mate_chrom as mate,
-           dm.mate_start as mateStart,
+           dm.mate_start - 1 as mateStart,
            dm.mate_id as matePos,
            dm.ref as reference,
            dm.alt as alternate,
            dm.allele_count as alleleCount,
            dm.allele_num as callCount,
            dm.allele_freq as frequency,
-           dm.mate_start as "end",
+           dm.mate_start - 1 as "end",
            'BND' as variantType
      FROM data.mate AS dm
       JOIN beacon.available_datasets as av