Use coverage for assessing data quality

halfpipe/models.py

@@ -277,13 +277,14 @@ class PreAlignmentQC(Step):
     def get_score(self):
         """Returns score consisting of read count and filtering percentage."""
 
-        # We consider more that 200 mio reads to be perfect
+        # Effective coverage
         READ_COUNT_SCORING = (
-            (200, 1),
-            (100, 0.9),
-            (50, 0.8),
-            (10, 0.6),
-            (0, 0.5),
+            (5, 1),
+            (4, 0.9),
+            (3, 0.8),
+            (2, 0.7),
+            (0.5, 0.5),
+            (0, 0),
         )
 
         if self.fastp_report:
@@ -293,9 +294,9 @@ class PreAlignmentQC(Step):
             filter_rate = after / before
 
             # Readcount scoring
-            bases = self.fastp_report["summary"]["after_filtering"]["total_bases"]
+            bases = self.fastp_report["summary"]["before_filtering"]["total_bases"]
             estimated_coverage = (
-                bases // self.pipeline.get().assembly.organism.genome_size
+                bases / self.pipeline.get().assembly.organism.genome_size
             )
             for v, r in READ_COUNT_SCORING:
                 if estimated_coverage >= v: