make gap more costly than mutations in the aligner

cfe-lab · Jun 23, 2023 · 3cf1ee5 · 3cf1ee5
1 parent c0c41b5
commit 3cf1ee5
Show file tree

Hide file tree

Showing 3 changed files with 8 additions and 14 deletions.
diff --git a/intact/intact.py b/intact/intact.py
@@ -601,6 +601,10 @@ def small_frames(
 
     aligner = Align.PairwiseAligner()
     aligner.mode = 'global'
+    aligner.match_score = 2
+    aligner.mismatch_score = -1
+    aligner.open_gap_score = -1.5
+    aligner.extend_gap_score = -0.2
 
     def translate(seq, frame = 0, to_stop = False):
         for_translation = seq[frame:]

diff --git a/tests/expected-results-large/errors.json b/tests/expected-results-large/errors.json
@@ -33,7 +33,7 @@
         {
             "sequence_name": "KX505501.1",
             "error": "FrameshiftInOrf",
-            "message": "Smaller ORF tat_exon2 at 8375-8468 contains out of frame indels that impact 80 positions."
+            "message": "Smaller ORF tat_exon2 at 8375-8468 contains out of frame indels that impact 51 positions."
         },
         {
             "sequence_name": "KX505501.1",
@@ -94,11 +94,6 @@
             "error": "InternalStopInOrf",
             "message": "Smaller ORF tat_exon1 at 5829-6044 contains an internal stop codon"
         },
-        {
-            "sequence_name": "MN692074",
-            "error": "FrameshiftInOrf",
-            "message": "Smaller ORF rev_exon1 at 5968-6044 contains out of frame indels that impact 51 positions."
-        },
         {
             "sequence_name": "MN692074",
             "error": "InsertionInOrf",
@@ -107,7 +102,7 @@
         {
             "sequence_name": "MN692074",
             "error": "FrameshiftInOrf",
-            "message": "Smaller ORF tat_exon2 at 8375-8468 contains out of frame indels that impact 81 positions."
+            "message": "Smaller ORF tat_exon2 at 8375-8468 contains out of frame indels that impact 85 positions."
         },
         {
             "sequence_name": "MN692074",

diff --git a/tests/expected-results-small/errors.json b/tests/expected-results-small/errors.json
@@ -33,7 +33,7 @@
         {
             "sequence_name": "KX505501.1",
             "error": "FrameshiftInOrf",
-            "message": "Smaller ORF tat_exon2 at 8375-8468 contains out of frame indels that impact 80 positions."
+            "message": "Smaller ORF tat_exon2 at 8375-8468 contains out of frame indels that impact 51 positions."
         },
         {
             "sequence_name": "KX505501.1",
@@ -94,11 +94,6 @@
             "error": "InternalStopInOrf",
             "message": "Smaller ORF tat_exon1 at 5829-6044 contains an internal stop codon"
         },
-        {
-            "sequence_name": "MN692074",
-            "error": "FrameshiftInOrf",
-            "message": "Smaller ORF rev_exon1 at 5968-6044 contains out of frame indels that impact 51 positions."
-        },
         {
             "sequence_name": "MN692074",
             "error": "InsertionInOrf",
@@ -107,7 +102,7 @@
         {
             "sequence_name": "MN692074",
             "error": "FrameshiftInOrf",
-            "message": "Smaller ORF tat_exon2 at 8375-8468 contains out of frame indels that impact 81 positions."
+            "message": "Smaller ORF tat_exon2 at 8375-8468 contains out of frame indels that impact 85 positions."
         },
         {
             "sequence_name": "MN692074",