dynamic interval calculation for accuracy

pdimens · Mar 3, 2025 · 17e4784 · 17e4784
1 parent 3b9466d
commit 17e4784
Show file tree

Hide file tree

Showing 4 changed files with 4 additions and 10 deletions.
diff --git a/harpy/bin/molecule_coverage.py b/harpy/bin/molecule_coverage.py
@@ -69,7 +69,7 @@ def which_overlap(start: int, end: int, binlist: list):
 def print_depth_counts(contig, counter_obj, intervals):
     """Print the Counter object to stdout"""
     for idx,int_bin in enumerate(intervals):
-        sys.stdout.write(f"{contig}\t{int_bin.start}\t{int_bin.stop}\t{counter_obj[idx]/args.window}\n")
+        sys.stdout.write(f"{contig}\t{int_bin.start}\t{int_bin.stop}\t{counter_obj[idx]/(int_bin.end - int_bin.start)}\n")
 
 with gzip.open(args.statsfile, "rt") as statsfile:
     aln_ranges = []

diff --git a/harpy/snakefiles/align_bwa.smk b/harpy/snakefiles/align_bwa.smk
@@ -211,12 +211,10 @@ rule alignment_coverage:
         bed = outdir + "/reports/data/coverage/coverage.bed"
     output: 
         outdir + "/reports/data/coverage/{sample}.cov.gz"
-    params:
-        windowsize
     container:
         None
     shell:
-        "samtools bedcov -c {input.bed} {input.bam} | awk '{{ $5 = $5 / {params}; print }}' | gzip > {output}"
+        "samtools bedcov -c {input.bed} {input.bam} | awk '{{ $5 = $5 / ($3 - $2); print }}' | gzip > {output}"
 
 rule report_config:
     input:

diff --git a/harpy/snakefiles/align_ema.smk b/harpy/snakefiles/align_ema.smk
@@ -279,12 +279,10 @@ rule alignment_coverage:
         bed = outdir + "/reports/data/coverage/coverage.bed"
     output: 
         outdir + "/reports/data/coverage/{sample}.cov.gz"
-    params:
-        windowsize
     container:
         None
     shell:
-        "samtools bedcov -c {input.bed} {input.bam} | awk '{{ $5 = $5 / {params}; print }}' | gzip > {output}"
+        "samtools bedcov -c {input.bed} {input.bam} | awk '{{ $5 = $5 / ($3 - $2); print }}' | gzip > {output}"
 
 rule barcode_stats:
     input:

diff --git a/harpy/snakefiles/align_strobealign.smk b/harpy/snakefiles/align_strobealign.smk
@@ -202,12 +202,10 @@ rule alignment_coverage:
         bed = outdir + "/reports/data/coverage/coverage.bed"
     output: 
         outdir + "/reports/data/coverage/{sample}.cov.gz"
-    params:
-        windowsize
     container:
         None
     shell:
-        "samtools bedcov -c {input.bed} {input.bam} | awk '{{ $5 = $5 / {params}; print }}' | gzip > {output}"
+        "samtools bedcov -c {input.bed} {input.bam} | awk '{{ $5 = $5 / ($3 - $2); print }}' | gzip > {output}"
 
 rule report_config:
     input: