Don't count g2p mapped reads twice, for #442.

micall/core/cascade_report.py

@@ -26,10 +26,17 @@ def parse_args():
 
 
 class CascadeReport:
-    def __init__(self, cascade_csv):
+    def __init__(self, cascade_csv, is_g2p_remapped=False):
+        """ Initialize a report object.
+
+        :param cascade_csv: an open CSV file to write to
+        :param is_g2p_remapped: True if the G2P reads get mapped again in the
+            remap step, so they shouldn't be included in the demultiplexed count
+        """
         self.g2p_summary_csv = self.remap_counts_csv = self.aligned_csv = None
         self.counts = None
         self.cascade_csv = cascade_csv
+        self.is_g2p_remapped = is_g2p_remapped
 
     def generate(self):
         field_names = ['demultiplexed',
@@ -57,7 +64,8 @@ def read_g2p(self):
         reader = DictReader(self.g2p_summary_csv)
         row = next(reader)
         mapped = int(row['mapped'])
-        self.counts['demultiplexed'] += mapped
+        if not self.is_g2p_remapped:
+            self.counts['demultiplexed'] += mapped
         self.counts['v3loop'] = mapped
         self.counts['g2p'] += int(row['valid'])
 

micall/drivers/sample.py

@@ -123,15 +123,16 @@ def process(self,
                  summary_file=read_summary,
                  use_gzip=use_gzip)
 
-        logger.info('Running merge_for_entropy on %s.', self)
-        with open(self.read_entropy_csv, 'w') as read_entropy_csv:
-            merge_for_entropy(self.trimmed1_fastq,
-                              self.trimmed2_fastq,
-                              read_entropy_csv,
-                              scratch_path)
+        if use_denovo:
+            logger.info('Running merge_for_entropy on %s.', self)
+            with open(self.read_entropy_csv, 'w') as read_entropy_csv:
+                merge_for_entropy(self.trimmed1_fastq,
+                                  self.trimmed2_fastq,
+                                  read_entropy_csv,
+                                  scratch_path)
 
-        write_merge_lengths_plot(self.read_entropy_csv,
-                                 self.merge_lengths_svg)
+            write_merge_lengths_plot(self.read_entropy_csv,
+                                     self.merge_lengths_svg)
 
         logger.info('Running fastq_g2p on %s.', self)
         with open(self.trimmed1_fastq) as fastq1, \
@@ -247,7 +248,8 @@ def process(self,
                 open(self.remap_counts_csv) as remap_counts_csv, \
                 open(self.aligned_csv) as aligned_csv, \
                 open(self.cascade_csv, 'w') as cascade_csv:
-            cascade_report = CascadeReport(cascade_csv)
+            cascade_report = CascadeReport(cascade_csv,
+                                           is_g2p_remapped=use_denovo)
             cascade_report.g2p_summary_csv = g2p_summary_csv
             cascade_report.remap_counts_csv = remap_counts_csv
             cascade_report.aligned_csv = aligned_csv

micall/tests/test_cascade_report.py

@@ -51,6 +51,30 @@ def test_remap_inputs(self):
 
         self.assertEqual(expected_report, report.cascade_csv.getvalue())
 
+    def test_g2p_inputs_remapped(self):
+        report = CascadeReport(StringIO(), is_g2p_remapped=True)
+        report.g2p_summary_csv = StringIO("""\
+mapped,valid,X4calls,X4pct,final,validpct
+100,90,0,0.00,,90.00
+""")
+        report.remap_counts_csv = StringIO("""\
+type,count,ignored
+raw,300,x
+prelim R1-seed,200,99
+prelim *,100,
+remap-1 R1-seed,220,
+remap-final R1-seed,240,
+unmapped,60,
+""")
+        expected_report = """\
+demultiplexed,v3loop,g2p,prelim_map,remap,aligned
+150,100,90,100,120,0
+"""
+
+        report.generate()
+
+        self.assertEqual(expected_report, report.cascade_csv.getvalue())
+
     def test_remap_inputs_two_seeds(self):
         report = CascadeReport(StringIO())
         report.remap_counts_csv = StringIO("""\