feat: Mashmap (#485)

* [fix] (template): Missing code in wrappers' doc. Error #187

* mashmap initial commit

* refList parameters included in tests

* fix fasta missing extension

* passing tests

* cleaning

* comment about relist

* formatting

* miss-renaming files

* miss-renaming files

* cleaning

* meta corrections : #485 (comment)

* Gzipped fasta support + query list as text file

* Gzipped support

* Gzipped support

* Documentation update #485 (comment)

Co-authored-by: tdayris <tdayris@gustaveroussy.fr>

bio/mashmap/environment.yaml

@@ -0,0 +1,8 @@
+channels:
+  - bioconda
+  - conda-forge
+  - defaults
+dependencies:
+  - mashmap =2.0
+  - gsl =2.7
+  - gzip =1.11

bio/mashmap/meta.yaml

@@ -0,0 +1,15 @@
+name: MashMap
+description: Compute local alignment boundaries between long DNA sequences with MashMap
+url: https://github.com/marbl/MashMap
+authors:
+  - Thibault Dayris
+input:
+  - ref: Path to reference file
+  - query: Path to query file (fasta, fastq)
+output:
+  - Path to the alignment file
+params:
+  - extra: Optional parameters for MashMap
+notes: |
+  * `input.ref` may be either a path to a fasta file or a text file containing a list of paths to several fasta files.
+  * `input.query` may be either a path to a fastq file or a text file containing a list of paths to several fastq files.

bio/mashmap/test/Snakefile

@@ -0,0 +1,13 @@
+rule test_mashmap:
+    input:
+        ref="reference.fasta.gz",  # This can be a txt file with a path to a fasta-file per line
+        query="read.fasta.gz",
+    output:
+        "mashmap.out",
+    threads: 2
+    params:
+        extra="-s 1000 --pi 99",
+    log:
+        "logs/mashmap.log",
+    wrapper:
+        "master/bio/mashmap"

bio/mashmap/test/Snakefile_reflist.smk

@@ -0,0 +1,12 @@
+rule test_mashmap_reflist:
+    input:
+        ref="reference.txt",
+        query="read.fasta.gz",
+    output:
+        "mashmap.out",
+    params:
+        extra="-s 500 --pi 99",
+    log:
+        "logs/mashmap.log",
+    wrapper:
+        "master/bio/mashmap"

bio/mashmap/test/reference.txt

@@ -0,0 +1 @@
+reference.fasta.gz

bio/mashmap/wrapper.py

@@ -0,0 +1,49 @@
+#!/usr/bin/python3.8
+# coding: utf-8
+
+""" Snakemake wrapper for MashMap """
+
+__author__ = "Thibault Dayris"
+__copyright__ = "Copyright 2022, Thibault Dayris"
+__email__ = "thibault.dayris@gustaveroussy.fr"
+__license__ = "MIT"
+
+from snakemake.shell import shell
+
+log = snakemake.log_fmt_shell(stdout=True, stderr=True)
+extra = snakemake.params.get("extra", "")
+max_threads = snakemake.threads
+
+# Handling input file types (either a fasta file, or a text file with a list of paths to fasta files)
+ref = snakemake.input["ref"]
+if ref.endswith(".txt"):
+    ref = f"--refList {ref}"
+elif ref.endswith(".gz"):
+    ref = f"--ref <( gzip --decompress --stdout {ref} )"
+    max_threads -= 1
+else:
+    ref = f"--ref {ref}"
+
+if max_threads < 1:
+    raise ValueError(
+        "Reference fasta on-the-fly g-unzipping consumed one thread."
+        f" Please increase the number of available threads by {1 - max_threads}."
+    )
+
+
+# Handling query file format (either a fastq file or a text file with a list of fastq files)
+query = snakemake.input["query"]
+if query.endswith(".txt"):
+    query = f"--queryList {query}"
+else:
+    query = f"--query {query}"
+
+shell(
+    "mashmap "
+    "{ref} "
+    "{query} "
+    "--output {snakemake.output} "
+    "--threads {snakemake.threads} "
+    "{extra} "
+    "{log}"
+)

test.py

@@ -130,6 +130,18 @@ def run(wrapper, cmd, check_log=None):
             os.chdir(origdir)
 
 
+@skip_if_not_modified
+def test_mashmap():
+    run(
+        "bio/mashmap",
+        ["snakemake", "--cores", "2", "mashmap.out", "--use-conda", "-F"]
+    )
+
+    run(
+        "bio/mashmap",
+        ["snakemake", "--cores", "2", "mashmap.out", "--use-conda", "-F", "-s", "Snakefile_reflist.smk"]
+    )
+
 @skip_if_not_modified
 def test_rbt_csvreport():
     run(