refine oversized seed penalty

AmpliconSuite · May 24, 2023 · 0a8a2ff · 0a8a2ff
1 parent cf0f61d
commit 0a8a2ff
Show file tree

Hide file tree

Showing 3 changed files with 11 additions and 11 deletions.
diff --git a/PrepareAA.py b/PrepareAA.py
@@ -15,7 +15,7 @@
 import check_reference
 import cnv_prefilter
 
-__version__ = "0.1537.1"
+__version__ = "0.1537.2"
 
 PY3_PATH = "python3"  # updated by command-line arg if specified
 metadata_dict = {}  # stores the run metadata (bioinformatic metadata)

diff --git a/README.md b/README.md
@@ -6,7 +6,7 @@ Performs preliminary steps (alignment, seed detection, & seed filtering) require
 
 AmpliconSuite-pipeline supports hg19, GRCh37, GRCh38 (hg38), and mouse genome mm10 (GRCm38). The tool also supports analysis with a human-viral hybrid reference genome we provide, "GRCh38_viral", which can be used to detect oncoviral hybrid focal amplifications and ecDNA in cancers with oncoviral infections.
 
-**Current version: 0.1537.1**
+**Current version: 0.1537.2**
 
 [comment]: # (Versioning based on major_version.days_since_initial_commit.minor_version. Initial commit: March 5th, 2019)
 

diff --git a/cnv_prefilter.py b/cnv_prefilter.py
@@ -159,23 +159,23 @@ def prefilter_bed(bedfile, ref, centromere_dict, chr_sizes, cngain, outdir):
         init_cns = arm2cns[a]
         med_cn = compute_cn_median(init_cns, arm2lens[a])
         for x in init_cns:
-            # ignore seeds over 30 Mbp
+            long_seed_region_penalty_mult = 1.
+            # ignore CN segments over 30 Mbp
             if x[2] - x[1] > 30000000:
                 continue
 
-            ccg = cngain
+            # penalize segments over 20 Mbp
+            elif x[2] - x[1] > 20000000:
+                long_seed_region_penalty_mult = 2.0
+
             continuous_high_hits = continuous_high_region_ivald[x[0]][x[1]:x[2]]
             if continuous_high_hits:
-                long_seed_region_penalty_mult = 1.
                 for y in continuous_high_hits:
-                    if y.end - y.begin > 20000000:
-                        long_seed_region_penalty_mult = max(3.0, long_seed_region_penalty_mult)
-
-                    elif y.end - y.begin > 10000000:
+                    # penalize seeds that overlap a high-CN region of 10 Mbp or more
+                    if y.end - y.begin > 10000000:
                         long_seed_region_penalty_mult = max(1.5, long_seed_region_penalty_mult)
 
-                ccg*=long_seed_region_penalty_mult
-
+            ccg = cngain * long_seed_region_penalty_mult
             if x[3] > med_cn + ccg - 2:
                 cn_filt_entries.append(x)