feat: added Bellerophon wrapper (#529)

* Added Bellerophon

* Added sorting

* Fixed log

* Added sort doc

* Fixed tests

* fix env

Co-authored-by: Johannes Köster <johannes.koester@uni-due.de>

bio/bellerophon/environment.yaml

@@ -0,0 +1,8 @@
+channels:
+  - conda-forge
+  - bioconda
+  - nodefaults
+dependencies:
+  - bellerophon =1.0
+  - samtools =1.15
+  - snakemake-wrapper-utils =0.4

bio/bellerophon/meta.yaml

@@ -0,0 +1,14 @@
+name: Bellerophon
+description: Filter mapped reads where the mapping spans a junction, retaining the 5-prime read.
+url: https://github.com/davebx/bellerophon/
+authors:
+  - Filipe G. Vieira
+input:
+  - Forward reads (BAM format)
+  - Reverse reads (BAM format)
+output:
+  - SAM/BAM/CRAM file
+notes: |
+  * The `sort` param allows to enable sorting ('none', 'queryname' or 'coordinate').
+  * The `sort_extra` allows for extra arguments for samtools.
+  * The `extra` param allows for additional program arguments.

bio/bellerophon/test/Snakefile

@@ -0,0 +1,32 @@
+rule bellerophon_sam:
+    input:
+        fwd="test_1500_forward.bam",
+        rev="test_1500_reverse.bam",
+    output:
+        bam="out.sam",
+    log:
+        "logs/bellerophon.log",
+    params:
+        extra="--quality 20",
+        sorting="none",  # optional: Enable sorting. Possible values: 'none', 'queryname' or 'coordinate'
+        sort_extra="--no-PG",  # optional: extra arguments for samtools/picard
+    threads: 2
+    wrapper:
+        "master/bio/bellerophon"
+
+
+rule bellerophon_bam:
+    input:
+        fwd="test_1500_forward.bam",
+        rev="test_1500_reverse.bam",
+    output:
+        bam="out.bam",
+    log:
+        "logs/bellerophon.log",
+    params:
+        extra="--quality 20",
+        sorting="coordinate",  # optional: Enable sorting. Possible values: 'none', 'queryname' or 'coordinate'
+        sort_extra="--no-PG",  # optional: extra arguments for samtools/picard
+    threads: 2
+    wrapper:
+        "master/bio/bellerophon"

bio/bellerophon/wrapper.py

@@ -0,0 +1,43 @@
+__author__ = "Filipe G. Vieira"
+__copyright__ = "Copyright 2022, Filipe G. Vieira"
+__license__ = "MIT"
+
+
+from snakemake.shell import shell
+from snakemake_wrapper_utils.samtools import get_samtools_opts
+
+
+samtools_opts = get_samtools_opts(snakemake, parse_output=False)
+log = snakemake.log_fmt_shell(stdout=False, stderr=True)
+extra = snakemake.params.get("extra", "")
+sort = snakemake.params.get("sorting", "none")
+sort_extra = snakemake.params.get("sort_extra", "")
+
+
+pipe_cmd = ""
+# Determine which pipe command to use for converting to bam or sorting.
+if sort == "none":
+    # Simply convert to output format using samtools view.
+    pipe_cmd = f"| samtools view -h {sort_extra} {samtools_opts}"
+elif sort in ["coordinate", "queryname"]:
+    # Add name flag if needed.
+    if sort == "queryname":
+        sort_extra += " -n"
+
+    # Sort alignments.
+    pipe_cmd = f"| samtools sort {sort_extra} {samtools_opts}"
+else:
+    raise ValueError(f"Unexpected value for params.sort: {sort}")
+
+
+shell(
+    "(bellerophon"
+    " --threads {snakemake.threads}"
+    " --forward {snakemake.input.fwd}"
+    " --reverse {snakemake.input.rev}"
+    " {extra}"
+    " --output /dev/stdout"
+    " {pipe_cmd}"
+    " > {snakemake.output[0]}"
+    ") {log}"
+)

test.py

@@ -241,6 +241,19 @@ def test_genomescope():
     )
 
 
+@skip_if_not_modified
+def test_bellerophon():
+    run(
+        "bio/bellerophon",
+        ["snakemake", "--cores", "2", "out.sam", "--use-conda", "-F"],
+    )
+
+    run(
+        "bio/bellerophon",
+        ["snakemake", "--cores", "2", "out.bam", "--use-conda", "-F"],
+    )
+
+
 @skip_if_not_modified
 def test_mashmap():
     run(