Merge pull request #18 from eipm/BIO-1104-bioinformatics-add-pigz-to-…

…the-image Added pigz, optimized multi-stage, updated README
eipm · Dec 6, 2022 · 2dcbe28 · 2dcbe28
2 parents ab2df1f + d8cb88d
commit 2dcbe28
Show file tree

Hide file tree

Showing 3 changed files with 22 additions and 110 deletions.
diff --git a/.github/workflows/docker-publish.yml b/.github/workflows/docker-publish.yml
@@ -84,32 +84,4 @@ jobs:
 
           docker tag $IMAGE_NAME $IMAGE_ID:$VERSION
           docker push $IMAGE_ID:$VERSION
-
-      - name: Login to CGEN DockerHub
-        uses: docker/login-action@v1
-        with:
-          username: ${{ secrets.CGEN_DOCKER_HUB_USERNAME }}
-          password: ${{ secrets.CGEN_DOCKER_HUB_TOKEN }}
-
-      - name: Push image to CGEN Docker Hub
-        run: |
-          IMAGE_ID=cgen/$IMAGE_NAME
-
-          # Change all uppercase to lowercase
-          IMAGE_ID=$(echo $IMAGE_ID | tr '[A-Z]' '[a-z]')
-
-          # Strip git ref prefix from version
-          VERSION=$(echo "${{ github.ref }}" | sed -e 's,.*/\(.*\),\1,')
-
-          # Strip "v" prefix from tag name
-          [[ "${{ github.ref }}" == "refs/tags/"* ]] && VERSION=$(echo $VERSION | sed -e 's/^v//')
-
-          # Use Docker `latest` tag convention
-          [ "$VERSION" == "main" ] && VERSION=latest
-
-          echo IMAGE_ID=$IMAGE_ID
-          echo VERSION=$VERSION
-
-          docker tag $IMAGE_NAME $IMAGE_ID:$VERSION
-          docker push $IMAGE_ID:$VERSION
   
diff --git a/Dockerfile b/Dockerfile
@@ -20,29 +20,12 @@ RUN apt-get update \
 	&& apt-get upgrade -y --fix-missing \
 	&& apt-get install build-essential -y \
 	&& apt-get install -y \
-	vim \
-	emacs \
-	bedtools \
-	bcftools \
-	vcftools \
-	bwa \
 	libncurses5-dev \
 	libbz2-dev \
 	liblzma-dev \
 	python-htseq \
 	&& rm -rf /var/lib/apt/lists/*
 
-#===========================#
-# Install BEDTOOLS			#
-#===========================#
-# ENV BEDTOOLS_VERSION 2.27.1
-# ENV bedtools_dir /${PROGRAMS}/bedtools-${BEDTOOLS_VERSION}
-# RUN wget -O bedtools-${BEDTOOLS_VERSION}.tar.gz https://github.com/arq5x/bedtools2/releases/download/v${BEDTOOLS_VERSION}/bedtools-${BEDTOOLS_VERSION}.tar.gz \
-#  	&& tar zxf bedtools-${BEDTOOLS_VERSION}.tar.gz -C ${PROGRAMS} \
-#  	&& rm bedtools-${BEDTOOLS_VERSION}.tar.gz \
-#  	&& cd ${PROGRAMS}/bedtools2 \
-#  	&& make \
-# 	&& make install
 #===========================#
 # Install SAMTOOLS & HTSLIB #
 #===========================#
@@ -59,39 +42,6 @@ RUN wget -O samtools-${SAMTOOLS_VERSION}.tar.bz2 https://github.com/samtools/sam
 	&& cd htslib-${HTSLIB_VERSION} \
 	&& make \
 	&& make install
-# #===========================#
-# # Install BCFTOOLS          #
-# #===========================#
-# ENV BCFTOOLS_VERSION 1.8
-# ENV bcftools_dir /${PROGRAMS}/bcftools-${BCFTOOLS_VERSION}
-# RUN wget -O bcftools-${BCFTOOLS_VERSION}.tar.bz2 https://github.com/samtools/bcftools/releases/download/${BCFTOOLS_VERSION}/bcftools-${BCFTOOLS_VERSION}.tar.bz2 \
-# 	&& tar jxf bcftools-${BCFTOOLS_VERSION}.tar.bz2 -C ${PROGRAMS} \
-# 	&& rm bcftools-${BCFTOOLS_VERSION}.tar.bz2 \
-# 	&& cd ${bcftools_dir} \
-# 	&& make \
-# 	&& make install
-#===========================#
-# Install VCFTOOLS			#
-#===========================#
-# ENV VCFTOOLS_VERSION 0.1.15
-# ENV vcftools_dir ${PROGRAMS}/vcftools-${VCFTOOLS_VERSION}
-# RUN wget -O vcftools-${VCFTOOLS_VERSION}.tar.gz https://github.com/vcftools/vcftools/releases/download/v${VCFTOOLS_VERSION}/vcftools-${VCFTOOLS_VERSION}.tar.gz \
-# 	&& tar zxf vcftools-${VCFTOOLS_VERSION}.tar.gz -C ${PROGRAMS} \
-# 	&& rm vcftools-${VCFTOOLS_VERSION}.tar.gz \
-# 	&& cd ${vcftools_dir} \
-# 	&& ./configure --bindir=/usr/local/bin \
-# 	&& make \
-# 	&& make install
-#===========================#
-# Install BWA				#
-#===========================#
-# ENV BWA_VERSION 0.7.17
-# ENV bwa_dir /${PROGRAMS}/bwa-${BWA_VERSION}
-# RUN wget -O bwa-${BWA_VERSION}.tar.bz2 http://sourceforge.net/projects/bio-bwa/files/bwa-${BWA_VERSION}.tar.bz2 \
-# 	&& tar jxf bwa-${BWA_VERSION}.tar.bz2 -C /${PROGRAMS} \
-# 	&& rm bwa-${BWA_VERSION}.tar.bz2 \
-# 	&& cd ${bwa_dir} \
-# 	&& make -f Makefile
 
 #===========================#
 # Install PINDEL			#
@@ -103,21 +53,7 @@ RUN cd ${PROGRAMS} \
 	&& git fetch origin pull/64/head:fix \
 	&& git checkout fix \
 	&& ./INSTALL ${htslib_dir}
-
-## PINDEL version: version 0.2.5b6, 20150915 (downloaded Nov 10 2015)
-# https://github.com/genome/pindel/archive/v${PINDEL_VERSION}.tar.gz
-# ENV PINDEL_VERSION 0.2.5b6
-# ENV pindel_dir /${PROGRAMS}/pindel
-# RUN wget -O pindel-master.zip https://github.com/genome/pindel/archive/master.zip \
-# 	&& unzip pindel-master.zip \	
-# 	&& rm pindel-master.zip \
-# 	&& mv pindel-master pindel \
-# 	&& cd pindel \
-# && ./INSTALL ${htslib_dir}/htslib-${HTSLIB_VERSION}
-# && ./INSTALL /${PROGRAMS}/samtools-${SAMTOOLS_VERSION}/htslib-${HTSLIB_VERSION}
-# RUN ln -s    ${bwa_dir}/bwa /usr/local/bin/bwa \
-# 	&& ln -s ${pindel_dir}/pindel /usr/local/bin/pindel 
-RUN ln -s ${pindel_dir}/pindel /usr/local/bin/pindel 
+RUN cp ${pindel_dir}/pindel /usr/local/bin/pindel 
 
 #===========================#
 # Install STAR              #
@@ -128,8 +64,8 @@ RUN wget -O STAR-${STAR_VERSION}.tar.gz https://github.com/alexdobin/STAR/archiv
 	&& tar xzf STAR-${STAR_VERSION}.tar.gz -C ${PROGRAMS} \
 	&& rm STAR-${STAR_VERSION}.tar.gz \
 	&& cd ${star_dir}/source \
-	&& make STAR 
-RUN ln -s ${star_dir}/source/STAR /usr/local/bin/
+	&& make STAR \
+	&& cp STAR /usr/local/bin/
 RUN apt-get upgrade -y && apt-get -y clean all
 
 ## Multi-stage build
@@ -144,17 +80,17 @@ ENV APP_NAME="bioinformatics" \
 
 RUN apt-get update \
 	&& apt-get upgrade -y --fix-missing \
+	&& apt-get install -y \
+	vim \
+	emacs \
+	bedtools \
+	bcftools \
+	vcftools \
+	bwa \ 
+	pigz \
 	&& apt-get -y clean all
 
-COPY --from=rstudio /usr/local /usr/local
-COPY --from=rstudio /usr/lib /usr/lib
-COPY --from=rstudio /usr/lib64 /usr/lib64
-RUN true
-COPY --from=rstudio /usr/bin /usr/bin
-RUN true
-COPY --from=rstudio /${PROGRAMS}/STAR-${STAR_VERSION} /${PROGRAMS}/STAR-${STAR_VERSION}
-RUN true
-COPY --from=rstudio /${PROGRAMS}/pindel/pindel /${PROGRAMS}/pindel/pindel
+COPY --from=rstudio /usr/local/bin /usr/local/bin
 RUN true
 COPY --from=rstudio /${PROGRAMS}/samtools-${SAMTOOLS_VERSION} /${PROGRAMS}/samtools-${SAMTOOLS_VERSION}
 

diff --git a/README.md b/README.md
@@ -2,19 +2,20 @@
 
 This application provides some of the basic bioinformatics tools for development, debugging, and troubleshooting applications.
 
-[![Actions Status](https://github.com/eipm/bioinformatics/workflows/Docker/badge.svg)](https://github.com/eipm/bioinformatics/actions) [![Github](https://img.shields.io/badge/github-1.3.1-green?style=flat&logo=github)](https://github.com/eipm/bioinformatics) [![EIPM Docker Hub](https://img.shields.io/badge/EIPM%20docker%20hub-1.3.1-blue?style=flat&logo=docker)](https://hub.docker.com/repository/docker/eipm/bioinformatics) [![CGEN Docker Hub](https://img.shields.io/badge/CGEN%20docker%20hub-1.3.1-blue?style=flat&logo=docker)](https://hub.docker.com/repository/docker/cgen/bioinformatics) [![GitHub Container Registry](https://img.shields.io/badge/GitHub%20Container%20Registry-1.3.1-blue?style=flat&logo=docker)](https://github.com/orgs/eipm/packages/container/package/bioinformatics)
+[![Actions Status](https://github.com/eipm/bioinformatics/workflows/Docker/badge.svg)](https://github.com/eipm/bioinformatics/actions) [![Github](https://img.shields.io/badge/github-1.3.1-green?style=flat&logo=github)](https://github.com/eipm/bioinformatics) [![EIPM Docker Hub](https://img.shields.io/badge/EIPM%20docker%20hub-1.3.1-blue?style=flat&logo=docker)](https://hub.docker.com/repository/docker/eipm/bioinformatics) [![GitHub Container Registry](https://img.shields.io/badge/GitHub%20Container%20Registry-1.3.1-blue?style=flat&logo=docker)](https://github.com/orgs/eipm/packages/container/package/bioinformatics)
 
 It includes:
 
-* R 4.0.2
-* bedtools (installed from distro with apt install: current version 2.27.1)
-* samtools & htslib 1.8
-* bcftools (installed from distro with apt install: current 1.10.2-3 using htslib 1.10.2-3)
+* R 4.2.2
+* bedtools (installed from distro with apt install: current version v2.30.0)
+* bcftools (installed from distro with apt install: current 1.13 (using htslib 1.13+ds)
 * vcftools (installed from distro with apt install: current 0.1.16)
 * bwa (installed from distro with apt install: current 0.7.17-r1188)
+* samtools (1.9 (using htslib 1.9))
 * pindel (latest:0.2.5b9, 20160729)
+* STAR (2.7.6a)
 
-The installed tools are the latest available as of Aug 18, 2020.
+The installed tools are the latest available as of Dec 5, 2022.
 
 **Note**: *pindel* includes a fix from a non-merged branch (see [Dockerfile](./Dockerfile)).
 
@@ -49,3 +50,6 @@ docker run --rm -it --name bioinfo -u $(id -u):$(id -g) -v /path/to/local/folder
 ### Transform BAM
 
 This image also include an utility called `transformBAM.sh` that will transform a BAM file by replacing IDs in the header. It will warn if changes need to occur in the content of the BAM file, e.g. the RG group. To use it, simply type `transformBAM.sh` and the usage will be displayed.
+
+### Combine pindel VCFs
+This utility `combine_pindel_vcfs.sh` takes multiple pindel results and merge them [TBDs]