feat: Deeptools bampefragmentsize

bio/deeptools/bampefragmentsize/environment.yaml

@@ -0,0 +1,6 @@
+channels:
+  - conda-forge
+  - bioconda
+  - nodefaults
+dependencies:
+  - deeptools =3.5.5

bio/deeptools/bampefragmentsize/meta.yaml

@@ -0,0 +1,34 @@
+name: deepTools computeMatrix
+url: https://deeptools.readthedocs.io/en/latest/content/tools/bamPEFragmentSize.html
+description: >
+  ``deepTools bamPEFragmentSize`` calculates the fragment sizes for read pairs given a BAM file from paired-end sequencing.
+  For usage information about ``deepTools bamPEFragmentSize``, please see the `documentation <https://deeptools.readthedocs.io/en/latest/content/tools/bamPEFragmentSize.html>`_.
+  For more information about ``deepTools``, also see the `source code <https://github.com/deeptools/deepTools>`_.
+
+    +----------------------------+-----------------------------+---------------------+-------------+
+    | bamPEFragmentSize          | Output format               | Name of output      | Recommended |
+    |                            |                             |                     |             |
+    |                            |                             | variable to be used | extension   |
+    +============================+=============================+=====================+=============+
+    | --histogram, -hist, -o     | gzipped matrix file         | hist                | ".png"      |
+    |                            |                             |                     |             |
+    |                            |                             | (required)          |             |
+    +----------------------------+-----------------------------+---------------------+-------------+
+    | --outRawFragmentLengths    | tab-separated table of      | raw                 | ".tab"      |
+    |                            |                             |                     |             |
+    |                            | fragments lengths and       |                     |             |
+    |                            |                             |                     |             |
+    |                            | occurances                  |                     |             |
+    +----------------------------+-----------------------------+---------------------+-------------+
+
+authors:
+  - Niek Wit
+input:
+  - bams: List of BAM files (.bam)
+  - blacklist: Optional BED file with regions to skip (.bed)
+output:
+  - hist: Fragment size histogram (.png)
+  - raw: Raw fragment lengths (.tab) (optional)
+params:
+  - title: Title of the plot
+  - extra: Optional parameters given to computeMatrix

bio/deeptools/bampefragmentsize/test/Snakefile

@@ -0,0 +1,25 @@
+rule deeptools_bampe_fragmentsize:
+    input:
+        # Input BAM file(s)
+        bams=["a.bam", "b.bam"],
+        # Optional blacklist file
+        # blacklist="",
+    output:
+        # Please note that -o/hist/--histogram and --outRawFragmentLengths are exclusively defined via output files.
+        # Usable output variables, their extensions and which option they implicitly call are listed here:
+        # https://snakemake-wrappers.readthedocs.io/en/stable/wrappers/deeptools/bamPEFragmentSize.html.
+        # Required
+        hist="results/histogram.png",
+        # Optional output files
+        raw="results/raw.tab",
+    log:
+        "logs/deeptools/bampe_fragmentsize.log",
+    threads: 4
+    params:
+        # Labels can be changed to anything
+        # If left empty, the sample name will be used 
+        # (without path and .bam extension)
+        labels="",  
+        extra="--logScale",
+    wrapper:
+        "master/bio/deeptools/bampefragmentsize"

bio/deeptools/bampefragmentsize/wrapper.py

@@ -0,0 +1,51 @@
+__author__ = "Niek Wit"
+__copyright__ = "Copyright 2025, Niek Wit"
+__email__ = "niekwit@gmail.com"
+__license__ = "MIT"
+
+import os
+from snakemake.shell import shell
+
+log = snakemake.log_fmt_shell(stdout=True, stderr=True)
+
+# Get input files
+bam_files = snakemake.input.get("bams", "")
+blacklist = snakemake.input.get("blacklist", "")
+if blacklist:
+    blacklist = f"--blackListFileName {blacklist}"
+
+# Get/create sample labels (remove .bam extension and dir)
+# If no labels are provided, use the basename of the bam file
+sample_label = snakemake.params.get("labels", "")
+if not sample_label:
+    sample_label = [os.path.basename(bam.replace(".bam", "")) for bam in bam_files]
+
+# Check if the number of labels is equal to the number of bam files
+assert len(sample_label) == len(
+    bam_files
+), "Number of labels must be equal to the number of bam files"
+
+out_file = snakemake.output.get("hist")
+
+# Check output format
+out_format = out_file.split(".")[-1]
+if not out_format in ["png", "pdf", "svg", "eps", "plotly"]:
+    raise ValueError(
+        "Output format must be either 'png', 'pdf', 'svg', 'eps', or 'plotly'"
+    )
+
+# Optional output
+out_raw = snakemake.output.get("raw", "")
+if out_raw:
+    optional_output = " --outRawFragmentLengths {out_tab} ".format(out_tab=out_raw)
+
+# Parameters
+extra = snakemake.params.get("extra", "")
+
+shell(
+    "bamPEFragmentSize "
+    "--numberOfProcessors {snakemake.threads} "
+    "-b {bam_files} "
+    "-o {out_file} "
+    "{blacklist} {optional_output} {extra} {log}"
+)

test_wrappers.py

@@ -2777,6 +2777,13 @@ def test_deeptools_bamcoverage(run):
     )
 
 
+def test_deeptools_bampe_fragmentsize(run):
+    run(
+        "bio/deeptools/bampefragmentsize",
+        ["snakemake", "--cores", "1", "results/histogram.png", "--use-conda", "-F"],
+    )
+
+
 def test_deeptools_multibigwigsummary(run):
     run(
         "bio/deeptools/multibigwigsummary",