Merge pull request #272 from LCR-BCCRC/module/fishhook/1.0

Module/fishhook/1.0

envs/fishhook/fishhook.yaml

@@ -0,0 +1,177 @@
+name: fishhook
+channels:
+  - conda-forge
+  - defaults
+dependencies:
+  - _libgcc_mutex=0.1=conda_forge
+  - _openmp_mutex=4.5=2_gnu
+  - _r-mutex=1.0.0=anacondar_1
+  - _sysroot_linux-64_curr_repodata_hack=3=haa98f57_10
+  - binutils_impl_linux-64=2.38=h2a08ee3_1
+  - binutils_linux-64=2.38.0=hc2dff05_0
+  - blas=1.0=openblas
+  - bwidget=1.9.11=1
+  - bzip2=1.0.8=h7b6447c_0
+  - c-ares=1.19.0=h5eee18b_0
+  - ca-certificates=2023.05.30=h06a4308_0
+  - cairo=1.16.0=hb05425b_4
+  - curl=7.88.1=h5eee18b_1
+  - expat=2.4.9=h6a678d5_0
+  - fontconfig=2.14.1=h4c34cd2_2
+  - freetype=2.12.1=h4a9f257_0
+  - fribidi=1.0.10=h7b6447c_0
+  - gcc_impl_linux-64=11.2.0=h1234567_1
+  - gcc_linux-64=11.2.0=h5c386dc_0
+  - gfortran_impl_linux-64=11.2.0=h1234567_1
+  - gfortran_linux-64=11.2.0=hc2dff05_0
+  - glib=2.69.1=he621ea3_2
+  - graphite2=1.3.14=h295c915_1
+  - gxx_impl_linux-64=11.2.0=h1234567_1
+  - gxx_linux-64=11.2.0=hc2dff05_0
+  - harfbuzz=4.3.0=hf52aaf7_1
+  - icu=58.2=he6710b0_3
+  - jpeg=9e=h5eee18b_1
+  - kernel-headers_linux-64=3.10.0=h57e8cba_10
+  - krb5=1.20.1=h143b758_1
+  - ld_impl_linux-64=2.38=h1181459_1
+  - lerc=3.0=h295c915_0
+  - libcurl=7.88.1=h251f7ec_1
+  - libdeflate=1.17=h5eee18b_0
+  - libedit=3.1.20221030=h5eee18b_0
+  - libev=4.33=h7f8727e_1
+  - libffi=3.4.4=h6a678d5_0
+  - libgcc-devel_linux-64=11.2.0=h1234567_1
+  - libgcc-ng=13.1.0=he5830b7_0
+  - libgfortran-ng=11.2.0=h00389a5_1
+  - libgfortran5=11.2.0=h1234567_1
+  - libgomp=13.1.0=he5830b7_0
+  - libnghttp2=1.52.0=h2d74bed_1
+  - libopenblas=0.3.21=h043d6bf_0
+  - libpng=1.6.39=h5eee18b_0
+  - libssh2=1.10.0=hdbd6064_2
+  - libstdcxx-devel_linux-64=11.2.0=h1234567_1
+  - libstdcxx-ng=13.1.0=hfd8a6a1_0
+  - libtiff=4.5.0=h6a678d5_2
+  - libuuid=1.41.5=h5eee18b_0
+  - libwebp-base=1.2.4=h5eee18b_1
+  - libxcb=1.15=h7f8727e_0
+  - libxml2=2.10.3=hcbfbd50_0
+  - lz4-c=1.9.4=h6a678d5_0
+  - make=4.2.1=h1bed415_1
+  - ncurses=6.4=h6a678d5_0
+  - openssl=3.0.9=h7f8727e_0
+  - pango=1.50.7=h05da053_0
+  - pcre=8.45=h295c915_0
+  - pcre2=10.37=he7ceb23_1
+  - pip=23.1.2=py311h06a4308_0
+  - pixman=0.40.0=h7f8727e_1
+  - python=3.11.3=h955ad1f_1
+  - r-askpass=1.1=r42h76d94ec_0
+  - r-assertthat=0.2.1=r42hc72bb7e_4
+  - r-base=4.2.0=h1ae530e_0
+  - r-base64enc=0.1_3=r42h57805ef_1006
+  - r-brew=1.0_8=r42hc72bb7e_2
+  - r-brio=1.1.3=r42h76d94ec_0
+  - r-bslib=0.5.0=r42hc72bb7e_1
+  - r-cachem=1.0.6=r42h76d94ec_0
+  - r-callr=3.7.3=r42hc72bb7e_1
+  - r-cli=3.6.1=r42ha503ecb_1
+  - r-clipr=0.8.0=r42hc72bb7e_2
+  - r-commonmark=1.8.0=r42h76d94ec_0
+  - r-cpp11=0.4.3=r42hc72bb7e_1
+  - r-crayon=1.5.2=r42hc72bb7e_2
+  - r-credentials=1.3.2=r42hc72bb7e_2
+  - r-curl=4.3.2=r42h76d94ec_0
+  - r-desc=1.4.2=r42hc72bb7e_2
+  - r-devtools=2.4.5=r42hc72bb7e_1
+  - r-diffobj=0.3.5=r42h76d94ec_0
+  - r-digest=0.6.29=r42h884c59f_0
+  - r-downlit=0.4.2=r42hc72bb7e_2
+  - r-ellipsis=0.3.2=r42h76d94ec_0
+  - r-evaluate=0.21=r42hc72bb7e_1
+  - r-fansi=1.0.3=r42h76d94ec_0
+  - r-fastmap=1.1.0=r42h884c59f_0
+  - r-fontawesome=0.5.1=r42hc72bb7e_1
+  - r-fs=1.5.2=r42h884c59f_0
+  - r-gert=1.6.0=r42h76d94ec_0
+  - r-gh=1.4.0=r42hc72bb7e_1
+  - r-gitcreds=0.1.2=r42hc72bb7e_2
+  - r-glue=1.6.2=r42h76d94ec_0
+  - r-highr=0.10=r42hc72bb7e_1
+  - r-htmltools=0.5.5=r42ha503ecb_1
+  - r-htmlwidgets=1.6.2=r42hc72bb7e_1
+  - r-httpuv=1.6.11=r42ha503ecb_1
+  - r-httr=1.4.6=r42hc72bb7e_1
+  - r-httr2=0.2.3=r42hc72bb7e_1
+  - r-ini=0.3.1=r42hc72bb7e_1005
+  - r-jquerylib=0.1.4=r42hc72bb7e_2
+  - r-jsonlite=1.8.0=r42h76d94ec_0
+  - r-knitr=1.39=r42h6115d3f_0
+  - r-later=1.3.1=r42ha503ecb_1
+  - r-lifecycle=1.0.1=r42h142f84f_0
+  - r-magrittr=2.0.3=r42h76d94ec_0
+  - r-memoise=2.0.1=r42hc72bb7e_2
+  - r-mime=0.12=r42h76d94ec_0
+  - r-miniui=0.1.1.1=r42hc72bb7e_1004
+  - r-openssl=2.0.2=r42h76d94ec_0
+  - r-pillar=1.9.0=r42hc72bb7e_1
+  - r-pkgbuild=1.4.2=r42hc72bb7e_0
+  - r-pkgconfig=2.0.3=r42hc72bb7e_3
+  - r-pkgdown=2.0.7=r42hc72bb7e_1
+  - r-pkgload=1.3.2=r42hc72bb7e_1
+  - r-praise=1.0.0=r42hc72bb7e_1007
+  - r-prettyunits=1.1.1=r42hc72bb7e_3
+  - r-processx=3.7.0=r42h06615bd_1
+  - r-profvis=0.3.8=r42h57805ef_0
+  - r-promises=1.2.0.1=r42ha503ecb_2
+  - r-ps=1.7.0=r42h76d94ec_0
+  - r-purrr=0.3.4=r42h76d94ec_0
+  - r-r6=2.5.1=r42hc72bb7e_2
+  - r-ragg=1.2.5=r42hd65d3ba_0
+  - r-rappdirs=0.3.3=r42h76d94ec_0
+  - r-rcmdcheck=1.4.0=r42h785f33e_2
+  - r-rcpp=1.0.10=r42ha503ecb_1
+  - r-rematch2=2.1.2=r42hc72bb7e_3
+  - r-remotes=2.4.2=r42hc72bb7e_2
+  - r-rlang=1.1.1=r42ha503ecb_1
+  - r-rmarkdown=2.14=r42h6115d3f_0
+  - r-roxygen2=7.2.3=r42ha503ecb_1
+  - r-rprojroot=2.0.3=r42hc72bb7e_0
+  - r-rstudioapi=0.14=r42hc72bb7e_2
+  - r-rversions=2.1.2=r42hc72bb7e_2
+  - r-sass=0.4.6=r42ha503ecb_1
+  - r-sessioninfo=1.2.2=r42hc72bb7e_2
+  - r-shiny=1.7.4=r42h785f33e_1
+  - r-sourcetools=0.1.7_1=r42ha503ecb_1
+  - r-stringi=1.7.6=r42h884c59f_0
+  - r-stringr=1.4.0=r42h6115d3f_0
+  - r-sys=3.4=r42h76d94ec_0
+  - r-systemfonts=1.0.4=r42haf97adc_2
+  - r-testthat=3.1.4=r42h884c59f_0
+  - r-textshaping=0.3.6=r42h884c59f_0
+  - r-tibble=3.1.7=r42h76d94ec_0
+  - r-tinytex=0.45=r42hc72bb7e_1
+  - r-urlchecker=1.0.1=r42hc72bb7e_2
+  - r-usethis=2.2.1=r42hc72bb7e_0
+  - r-utf8=1.2.2=r42h76d94ec_0
+  - r-vctrs=0.4.1=r42h884c59f_0
+  - r-waldo=0.5.1=r42hc72bb7e_1
+  - r-whisker=0.4.1=r42hc72bb7e_1
+  - r-withr=2.5.0=r42hc72bb7e_2
+  - r-xfun=0.31=r42h76d94ec_0
+  - r-xml2=1.3.3=r42h884c59f_0
+  - r-xopen=1.0.0=r42hc72bb7e_1005
+  - r-xtable=1.8_4=r42hc72bb7e_5
+  - r-yaml=2.3.5=r42h76d94ec_0
+  - r-zip=2.2.0=r42h76d94ec_0
+  - readline=8.2=h5eee18b_0
+  - setuptools=67.8.0=py311h06a4308_0
+  - sqlite=3.41.2=h5eee18b_0
+  - sysroot_linux-64=2.17=h57e8cba_10
+  - tk=8.6.12=h1ccaba5_0
+  - tktable=2.10=h14c3975_0
+  - wheel=0.38.4=py311h06a4308_0
+  - xz=5.4.2=h5eee18b_0
+  - zlib=1.2.13=h5eee18b_0
+  - zstd=1.5.5=hc292b87_0
+prefix: /home/jacky/miniconda3/envs/fishhook

modules/fishhook/1.0/config/default.yaml

@@ -0,0 +1,41 @@
+lcr-modules:
+
+    fishhook:
+        inputs:
+            # Available wildcards: {seq_type}
+            master_maf: "_UPDATE_"
+            sample_sets: "_UPDATE_"
+
+        sample_set: ["_UPDATE_"]
+        seq_types: ["genome", "capture"]
+        include_non_coding: True
+        prepare_mafs: "{SCRIPTSDIR}/generate_smg_inputs/1.0/generate_smg_inputs.R"
+        src_dir: "{MODSDIR}/src/"
+
+        options:
+            include_silent_mutation: False #Include silent mutations in maf file
+            tiles_size: 1000 #This is the size of tiles you want to split the genome into, leave blank if you want to use a gene list as hypothese
+            gene_list: "" #If you do not provide a tile size, please provide a gene list in gtf format (e.g GENCODE v19 genes for GRCH37: gencode.v19.genes.gtf) for Fishhook to do a gene-centric analysis
+            gene_list_only_protein_coding: True #Only include protein coding genes in gene list
+            covariates: "" #please supply a coveriate file in bed format, leave it blank if you want to run FishHook without covariates
+
+        conda_envs:
+            prepare_mafs: "{REPODIR}/envs/gatk/gatkR.yaml"
+            fishhook: "{MODSDIR}/envs/fishhook.yaml"
+
+        threads:
+            fishhook: 1
+
+        resources:
+            fishhook: 
+                mem_mb: 16000
+
+        pairing_config:
+            genome:
+                run_paired_tumours: True
+                run_unpaired_tumours_with: "unmatched_normal"
+                run_paired_tumours_as_unpaired: False
+            capture:
+                run_paired_tumours: True
+                run_unpaired_tumours_with: "unmatched_normal"
+                run_paired_tumours_as_unpaired: False           

modules/fishhook/1.0/envs/fishhook.yaml

@@ -0,0 +1 @@
+../../../../envs/fishhook/fishhook.yaml

modules/fishhook/1.0/fishhook.smk

@@ -0,0 +1,171 @@
+#!/usr/bin/env snakemake
+
+
+##### ATTRIBUTION #####
+
+
+# Original Author:  Jacky Yiu
+# Module Author:    Jacky Yiu
+# Contributors:     N/A
+
+
+##### SETUP #####
+
+# Import package with useful functions for developing analysis modules
+import oncopipe as op
+
+# Check that the oncopipe dependency is up-to-date. Add all the following lines to any module that uses new features in oncopipe
+min_oncopipe_version="1.0.11"
+import pkg_resources
+try:
+    from packaging import version
+except ModuleNotFoundError:
+    sys.exit("The packaging module dependency is missing. Please install it ('pip install packaging') and ensure you are using the most up-to-date oncopipe version")
+
+# To avoid this we need to add the "packaging" module as a dependency for LCR-modules or oncopipe
+
+current_version = pkg_resources.get_distribution("oncopipe").version
+if version.parse(current_version) < version.parse(min_oncopipe_version):
+    print('\x1b[0;31;40m' + f'ERROR: oncopipe version installed: {current_version}' + '\x1b[0m')
+    print('\x1b[0;31;40m' + f"ERROR: This module requires oncopipe version >= {min_oncopipe_version}. Please update oncopipe in your environment" + '\x1b[0m')
+    sys.exit("Instructions for updating to the current version of oncopipe are available at https://lcr-modules.readthedocs.io/en/latest/ (use option 2)")
+
+# End of dependency checking section
+
+# Setup module and store module-specific configuration in `CFG`
+# `CFG` is a shortcut to `config["lcr-modules"]["fishhook"]`
+CFG = op.setup_module(
+    name = "fishhook",
+    version = "1.0",
+    subdirectories = ["inputs", "fishhook", "outputs"],
+)
+
+# Define rules to be run locally when using a compute cluster
+localrules:
+    _fishhook_input_maf,
+    _fishhook_input_subsets,
+    _fishhook_prepare_maf,
+    _install_fishhook,
+    _run_fishhook,
+    _fishhook_output_tsv,
+    _fishhook_all,
+
+
+##### RULES #####
+
+
+# Symlinks the input files into the module results directory (under '00-inputs/')
+rule _fishhook_input_maf:
+    input:
+        maf = CFG["inputs"]["master_maf"]
+    output:
+        maf = CFG["dirs"]["inputs"] + "maf/{seq_type}/input.maf"
+    run:
+        op.absolute_symlink(input.maf, output.maf)
+
+# Symlinks the input files into the module results directory (under '00-inputs/')
+rule _fishhook_input_subsets:
+    input:
+        sample_sets = CFG["inputs"]["sample_sets"]
+    output:
+        sample_sets = CFG["dirs"]["inputs"] + "sample_sets/sample_sets.tsv"
+    run:
+        op.absolute_symlink(input.sample_sets, output.sample_sets)
+
+# Prepare the maf file for the input to Fishhook
+rule _fishhook_prepare_maf:
+    input:
+        maf = expand(
+                    str(rules._fishhook_input_maf.output.maf),
+                    allow_missing=True,
+                    seq_type=CFG["seq_types"]
+                    ),
+        sample_sets = ancient(str(rules._fishhook_input_subsets.output.sample_sets))
+    output:
+        maf = temp(CFG["dirs"]["inputs"] + "maf/{sample_set}.maf"),
+        contents = CFG["dirs"]["inputs"] + "maf/{sample_set}.maf.content"
+    log:
+        stdout = CFG["logs"]["inputs"] + "{sample_set}/prepare_maf.stdout.log",
+        stderr = CFG["logs"]["inputs"] + "{sample_set}/prepare_maf.stderr.log"
+    conda:
+        CFG["conda_envs"]["prepare_mafs"]
+    params:
+        include_non_coding = str(CFG["include_non_coding"]).upper(),
+        script = CFG["prepare_mafs"]
+    shell:
+        op.as_one_line("""
+        Rscript {params.script}
+        {input.maf}
+        {input.sample_sets}
+        $(dirname {output.maf})/
+        {wildcards.sample_set}
+        FishHook
+        {params.include_non_coding}
+        > {log.stdout} 2> {log.stderr}
+        """)
+
+
+# Install fishhook
+# only available from github, not through conda/CRAN/Biocmanager
+rule _install_fishhook:
+    output:
+        complete = CFG["dirs"]["inputs"] + "fishhook_installed.success"
+    conda:
+        CFG["conda_envs"]["fishhook"]
+    log:
+        input = CFG["logs"]["inputs"] + "install_fishhook.log"
+    shell:
+        """
+        R -q -e 'devtools::install_github(c("jokergoo/ComplexHeatmap","mskilab/gTrack", "mskilab/fishHook"))' >> {log.input} &&
+        touch {output.complete}"""
+
+
+# Example variant calling rule (multi-threaded; must be run on compute server/cluster)
+rule _run_fishhook:
+    input:
+        fishhook = ancient(str(CFG["dirs"]["inputs"] + "fishhook_installed.success")),
+        maf = str(rules._fishhook_prepare_maf.output.maf)
+    output:
+        tsv = CFG["dirs"]["fishhook"] + "{sample_set}/fishhook.output.maf"
+    conda:
+        CFG["conda_envs"]["fishhook"]
+    threads:
+        CFG["threads"]["fishhook"]
+    resources:
+        **CFG["resources"]["fishhook"]     
+    params:
+        tiles_size = CFG["options"]["tiles_size"],
+        coveriate = CFG["options"]["covariates"],
+        include_silent = CFG["options"]["include_silent_mutation"],
+        gene_list = CFG["options"]["gene_list"],
+        gene_list_only_protein_coding = CFG["options"]["gene_list_only_protein_coding"]
+    script:
+        "scr/R/run_fishhook.R"
+
+
+# Symlinks the final output files into the module results directory (under '99-outputs/')
+rule _fishhook_output_tsv:
+    input:
+        tsv = str(rules._run_fishhook.output.tsv)
+    output:
+        tsv = CFG["dirs"]["outputs"] + "tsv/{sample_set}/{sample_set}.fishhook.tsv"
+    run:
+        op.relative_symlink(input.tsv, output.tsv, in_module= True)
+
+
+# Generates the target sentinels for each run, which generate the symlinks
+rule _fishhook_all:
+    input:
+        expand(
+            [
+                str(rules._fishhook_output_tsv.output.tsv),
+            ],
+            sample_set=CFG["sample_set"])
+
+
+##### CLEANUP #####
+
+
+# Perform some clean-up tasks, including storing the module-specific
+# configuration on disk and deleting the `CFG` variable
+op.cleanup_module(CFG)

modules/fishhook/1.0/schemas/base-1.0.yaml

@@ -0,0 +1 @@
+../../../../schemas/base/base-1.0.yaml