Fix build and add custom parameterization (#24)

* add battenberg.R for customization

* add max.rho to battenberg_bl_custom.R

* add max rho default

* add custom battenberg wrapper

* add max rho to wrapper

* add max rho to wrapper

* add max rho to wrapper

* use battenberg.R from v2.2.9 and add max.rho

* rename battenberg_wgs.R  battenberg_wgs_bl_custom.R

* resolve merge conflict again

* Add renv.lock file containining R package versions

* remove modify_reference_path.sh

* resolve merge conflict in Dockerfile

* resue ascat/battenberg version arg

* Update CHANGELOG.md

* Update changes

* commit to remove installer.R

* specify removal of files under /usr/local/src/

* verify checksum and set up workdir

* update CHANGELOG

* Update license year

---------

Co-authored-by: Mootor <mmootor@ip-0A125937.rhxrlfvjyzbupc03cc22jkch3c.xx.internal.cloudapp.net>

CHANGELOG.md

@@ -16,6 +16,19 @@ This project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.htm
 
 ---
 
+## [2.2.9-bl-parametrize-max-rho] - 2024-10-29
+### Added
+- Add `renv.lock` file containing R packages and their version
+- Add `battenberg_bl_custom.R`, a customized `battenberg.R` script that allows parameterization of `max_rho`
+- Add `battenberg_wgs_bl_custom.R`, a customized `battenberg_wgs.R` wrapper that allows parameterization of options
+
+### Changed
+- Reproduce R packages environment using `renv.lock` file
+- Install ASCAT and Battenberg using `R CMD INSTALL`
+
+### Removed
+- Remove `modify_reference_path.sh`
+
 ## [2.2.9] - 2023-06-27
 ### Added
 - Add `modify_reference_path.sh`

Dockerfile

@@ -1,4 +1,6 @@
 ARG MINIFORGE_VERSION=23.1.0-1
+ARG ASCAT_VERSION=3.1.2
+ARG BATTENBERG_VERSION=2.2.9
 
 FROM condaforge/mambaforge:${MINIFORGE_VERSION} AS builder
 
@@ -13,49 +15,73 @@ RUN mamba create -qy -p /usr/local \
     cancerit-allelecount==${ALLELECOUNT_VERSION} \
     impute2==${IMPUTE2_VERSION}
 
-FROM rocker/r-ver:4.4.1
+# Deploy the target tools into a base image
+FROM ubuntu:20.04
 COPY --from=builder /usr/local /usr/local
 
-RUN apt-get update \
-    && apt-get install -y --no-install-recommends \
-        libcurl4-openssl-dev \
-        libbz2-dev \
-        liblzma-dev \
-        libpng-dev \
-        libssl-dev \
-        libxml2-dev \
-        python3 \
-    && apt-get clean && rm -rf /var/lib/apt/lists/*
-
-# Main tool version
-ARG BATTENBERG_VERSION="2.2.9"
-
-# Dependency version or commit ID
-ARG ASCAT_VERSION=3.1.3
-ARG COPYNUMBER_VERSION="b404a4d"
-
-# GitHub repo link
-ARG ASCAT="VanLoo-lab/ascat/ASCAT@v${ASCAT_VERSION}"
-ARG COPYNUMBER="igordot/copynumber@${COPYNUMBER_VERSION}"
-ARG BATTENBERG="Wedge-lab/battenberg@v${BATTENBERG_VERSION}"
-
-# Install Package Dependency toolkit
-RUN R -e 'install.packages(c("argparse", "BiocManager", "pkgdepends", "optparse"))' && \
-    R -q -e 'BiocManager::install(c("ellipsis", "splines", "VariantAnnotation"))'
-
-# Install Battenberg
-COPY installer.R /usr/local/bin/installer.R
-RUN chmod +x /usr/local/bin/installer.R
-
-RUN Rscript /usr/local/bin/installer.R -d ${COPYNUMBER} ${ASCAT} ${BATTENBERG}
-
-# Modify paths to reference files
-COPY modify_reference_path.sh /usr/local/bin/modify_reference_path.sh
-RUN chmod +x /usr/local/bin/modify_reference_path.sh && \
-    bash /usr/local/bin/modify_reference_path.sh /usr/local/lib/R/site-library/Battenberg/example/battenberg_wgs.R /usr/local/bin/battenberg_wgs.R
-
-RUN ln -sf /usr/local/lib/R/site-library/Battenberg/example/filter_sv_brass.R /usr/local/bin/filter_sv_brass.R && \
-    ln -sf /usr/local/lib/R/site-library/Battenberg/example/battenberg_cleanup.sh /usr/local/bin/battenberg_cleanup.sh
+ARG DEBIAN_FRONTEND=noninteractive
+
+RUN apt-get update && \
+    apt-get install -y --no-install-recommends \
+    libxml2 \
+    libxml2-dev \
+    libcurl4-gnutls-dev \
+    build-essential \
+    libfontconfig1-dev \
+    libharfbuzz-dev \
+    libfribidi-dev \
+    libfreetype6-dev \
+    libpng-dev \
+    libtiff5-dev \
+    libjpeg-dev \
+    r-cran-rgl \
+    git \
+    libssl-dev \
+    r-cran-curl \
+    r-cran-devtools \
+    wget && \
+    apt-get clean && rm -rf /var/lib/apt/lists/*
+
+RUN R -q -e 'install.packages("renv")' && \
+    mkdir -p /usr/local/src
+
+COPY renv.lock /usr/local/renv.lock
+COPY battenberg_bl_custom.R /usr/local/src/
+COPY battenberg_wgs_bl_custom.R /usr/local/src/
+
+ARG ASCAT_VERSION
+ARG ASCAT_SHA512="a60e75405c3999c86d19e20d717eee9d1e1915e647e2626be5b5bb0d266a6f96535d9cd21ee7717d6fb04d4f76a5b78a1b45e3315bf823e480a8e27afdec364b"
+ARG BATTENBERG_VERSION
+ARG BATTENBERG_SHA512="a4784ca3e6523bd47b5a6d86c1e7ef5f0023371bd0b5bf3685440e3336333cddf3937f768910462c98eb5cb8b5cfd2c63db052c3186a3b1f0363e460ace521d4"
+
+WORKDIR /usr/local/src/
+
+RUN set -eux && \
+    # Ignore specific packages from `renv.lock` file
+    R -q -e 'renv::settings$ignored.packages(c("ASCAT", "Battenberg"))' && \
+    R -q -e 'renv::restore(lockfile = "/usr/local/renv.lock")' && \
+    # Install ASCAT
+    wget -q -O ascat-${ASCAT_VERSION}.tar.gz \
+        https://github.com/VanLoo-lab/ascat/archive/refs/tags/v${ASCAT_VERSION}.tar.gz && \
+    if echo "$ASCAT_SHA512" ascat-${ASCAT_VERSION}.tar.gz | sha512sum -c --quiet; \
+    then echo "ASCAT SHA512 checksum verified successfully!"; \
+    else echo "ASCAT SHA512 checksum verification failed. Downloaded file checksum does not match the SHA512 hash."; exit 1; \
+    fi && \
+    tar -xzf ascat-${ASCAT_VERSION}.tar.gz && \
+    R CMD INSTALL ascat-${ASCAT_VERSION}/ASCAT/ && \
+    # Instal Battenberg
+    wget -q -O battenberg-${BATTENBERG_VERSION}.tar.gz \
+        https://github.com/Wedge-lab/battenberg/archive/refs/tags/v${BATTENBERG_VERSION}.tar.gz && \
+    if echo "$BATTENBERG_SHA512" battenberg-${BATTENBERG_VERSION}.tar.gz | sha512sum -c --quiet; \
+    then echo "Battenberg SHA512 checksum verified successfully!"; \
+    else echo "Battenberg SHA512 checksum verification failed. Downloaded file checksum does not match the SHA512 hash."; exit 1; \
+    fi && \
+    tar -xzf battenberg-${BATTENBERG_VERSION}.tar.gz && \
+    cp battenberg_bl_custom.R battenberg-${BATTENBERG_VERSION}/R/battenberg.R && \
+    cp battenberg_wgs_bl_custom.R battenberg-${BATTENBERG_VERSION}/inst/example/battenberg_wgs.R && \
+    R CMD INSTALL battenberg-${BATTENBERG_VERSION}/ && \
+    # Cleanup
+    rm -rf /usr/local/src/*
 
 # Add a new user/group called bldocker
 RUN groupadd -g 500001 bldocker && \

README.md

@@ -1,27 +1,34 @@
 # docker-Battenberg
 This repository contains code for the whole genome sequencing subclonal copy number caller Battenberg, as described in [Nik-Zainal, Van Loo, Wedge, et al. (2012), Cell](https://www.ncbi.nlm.nih.gov/pubmed/22608083).
 
-It installs the release v2.2.9 of Battenberg and modifies the Battenberg resource paths for GRCh37 and GRCh38 based on how they are structured in the Boutros Lab cluster.
+It installs Battenberg v2.2.9 using
+- a custom battenberg.R script that allows parameterization of `max_rho`
+- a custom battenberg_wgs.R script that allows parameterization of all available options
 
-GRCh37 resources - `/hot/ref/tool-specific-input/Battenberg/download_202204/GRCh37/`
+## Resources
+GRCh37 resources - `/hot/resource/tool-specific-input/Battenberg/download_202204/GRCh37/`
 
 GRCh38 resources -
- - with `chr` name (default): `/hot/ref/tool-specific-input/Battenberg/download_202204/GRCh38/battenberg_ref_hg38_chr/`
- - without `chr` name: `/hot/ref/tool-specific-input/Battenberg/download_202204/GRCh38/battenberg_ref_hg38_non_chr/`
+ - with `chr` name (default): `/hot/resource/tool-specific-input/Battenberg/download_202204/GRCh38/battenberg_ref_hg38_chr/`
+ - without `chr` name: `/hot/resource/tool-specific-input/Battenberg/download_202204/GRCh38/battenberg_ref_hg38_non_chr/`
 
 This image can be found in docker-Battenberg's GitHub package page [here](https://github.com/uclahs-cds/docker-Battenberg/pkgs/container/battenberg).
 
 # Example Usage
 ```
 docker run --rm -u $(id -u):$(id -g) -w $(pwd) -v /hot/:/hot/ \
-    -v /hot/ref/tool-specific-input/Battenberg/download_202204/GRCh38/battenberg_ref_hg38_chr/:/opt/battenberg_reference/ \
-    battenberg:2.2.9 Rscript /usr/local/bin/battenberg_wgs.R \
+    -v /hot/resource/tool-specific-input/Battenberg/download_202204/GRCh38/battenberg_ref_hg38_chr/:/opt/battenberg_reference/ \
+    battenberg:2.2.9-bl-parametrize-max-rho Rscript /usr/local/lib/R/site-library/Battenberg/example/battenberg_wgs.R \
         -t ${tumor_sample_name} \
         -n ${normal_sample_name} \
         --tb ${tumor_bam} \
         --nb ${normal_bam} \
         -o ${sample_out_dir} \
-        --sex ${sample_sex}
+        --sex ${sample_sex} \
+        --min_ploidy 1.6 \
+        --max_ploidy 4.8 \
+        --min_rho 0.1 \
+        --max_rho 1.0
 ```
 
 # Documentation
@@ -65,9 +72,9 @@ Author: 'Mohammed Faizal Eeman Mootor', 'Ardalan Davarifar'
 
 docker-Battenberg is licensed under the GNU General Public License version 2. See the file LICENSE for the terms of the GNU GPL license.
 
-docker-Battenberg can be used to create a docker instance to use the Battenberg tool. 
+docker-Battenberg can be used to create a docker instance to use the Battenberg tool.
 
-Copyright (C) 2021-2023 University of California Los Angeles ("Boutros Lab") All rights reserved.
+Copyright (C) 2021-2024 University of California Los Angeles ("Boutros Lab") All rights reserved.
 
 This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version.