Merge pull request #988 from Kincekara/clair

add clair3

Program_Licenses.md

@@ -33,6 +33,7 @@ The licenses of the open-source software that is contained in these Docker image
 | Circlator | GNU GPLv3 | https://github.com/sanger-pathogens/circlator/blob/master/LICENSE |
 | Circos | GNU GPLv3 | https://circos.ca/ |
 | CirculoCov | GNU GPLv3 | https://github.com/erinyoung/CirculoCov/blob/main/LICENSE |
+| Clair3 | non-standard | https://github.com/HKU-BAL/Clair3/blob/main/LICENSE.md |
 | colorid | MIT | https://github.com/hcdenbakker/colorid/blob/master/LICENSE |
 | datasets-sars-cov-2 | Apache 2.0 | https://github.com/CDCgov/datasets-sars-cov-2/blob/master/LICENSE |
 | diamond | GNU GPLv3 | https://github.com/bbuchfink/diamond/blob/master/LICENSE |

README.md

@@ -136,6 +136,7 @@ To learn more about the docker pull rate limits and the open source software pro
 | [Circlator](https://hub.docker.com/r/staphb/circlator) <br/> [![docker pulls](https://badgen.net/docker/pulls/staphb/circlator)](https://hub.docker.com/r/staphb/circlator) | <ul><li>1.5.6</li><li>1.5.5</li></ul> | https://github.com/sanger-pathogens/circlator |
 | [Circos](https://hub.docker.com/r/staphb/circos) <br/> [![docker pulls](https://badgen.net/docker/pulls/staphb/circos)](https://hub.docker.com/r/staphb/circos) | <ul><li>[0.69-9](./circos/0.69.9/)</li></ul> | https://circos.ca/ |
 | [CirculoCov](https://hub.docker.com/r/staphb/circulocov) <br/> [![docker pulls](https://badgen.net/docker/pulls/staphb/circulocov)](https://hub.docker.com/r/staphb/circulocov) | <ul><li>[0.1.20240104](./circulocov/0.1.20240104/)</li></ul> | https://github.com/erinyoung/CirculoCov |
+| [Clair3](https://hub.docker.com/r/staphb/clair3) <br/> [![docker pulls](https://badgen.net/docker/pulls/staphb/clair3)](https://hub.docker.com/r/staphb/clair3) | <ul><li>[1.0.9](./clair3/1.0.9/)</li></ul> | https://github.com/HKU-BAL/Clair3 |
 | [Clustalo](https://hub.docker.com/r/staphb/clustalo) <br/> [![docker pulls](https://badgen.net/docker/pulls/staphb/clustalo)](https://hub.docker.com/r/staphb/clustalo) | <ul><li>1.2.4</li></ul> | http://www.clustal.org/omega/ |
 | [colorid](https://hub.docker.com/r/staphb/colorid) <br/> [![docker pulls](https://badgen.net/docker/pulls/staphb/colorid)](https://hub.docker.com/r/staphb/colorid) | <ul><li>0.1.4.3</li></ul> | https://github.com/hcdenbakker/colorid |
 | [cutshaw-report-env](https://hub.docker.com/r/staphb/cutshaw-report-env) <br/> [![docker pulls](https://badgen.net/docker/pulls/staphb/cutshaw-report-env)](https://hub.docker.com/r/staphb/cutshaw-report-env) | <ul><li>1.0.0</li></ul> | https://github.com/VADGS/CutShaw |

clair3/1.0.9/Dockerfile

@@ -0,0 +1,119 @@
+ARG CLAIR3_VER="1.0.9"
+
+FROM mambaorg/micromamba:1.5.8 as builder
+
+ARG CLAIR3_VER
+
+USER root
+
+WORKDIR /
+
+RUN apt-get update && apt-get install -y \
+    wget \
+    bzip2 \
+    make \
+    g++ \
+    libboost-graph-dev
+
+RUN micromamba install --name base -c conda-forge -c bioconda \
+    python=3.9.0 \
+    pypy3.6 \
+    tensorflow-cpu=2.8.0 \
+    pytables \
+    pigz \
+    cffi=1.14.4 \
+    parallel=20191122 \
+    zstd \
+    samtools=1.15.1 \
+    whatshap=1.7 \
+    xz \
+    zlib \
+    bzip2 \
+    automake \
+    curl &&\
+    micromamba clean -a -y
+
+ENV PATH="$PATH:/opt/conda/bin/"
+
+ARG MAMBA_DOCKERFILE_ACTIVATE=1
+
+RUN pypy3 -m ensurepip && \
+    pypy3 -m pip install mpmath==1.2.1 &&\
+    pip install tensorflow-addons
+
+RUN wget https://github.com/HKU-BAL/Clair3/archive/refs/tags/v${CLAIR3_VER}.tar.gz &&\
+    tar -xvf v${CLAIR3_VER}.tar.gz &&\
+    cd Clair3-${CLAIR3_VER}/preprocess/realign/ &&\
+    g++ -std=c++14 -O1 -shared -fPIC -o realigner ssw_cpp.cpp ssw.c realigner.cpp && \
+    g++ -std=c++11 -shared -fPIC -o debruijn_graph -O3 debruijn_graph.cpp &&\
+    cd ../.. &&\
+    make &&\
+    mkdir /clair3 &&\
+    cp -rv clair3 preprocess postprocess scripts shared /clair3 &&\
+    cp clair3.py run_clair3.sh /clair3 &&\
+    cp longphase libclair3* /clair3 &&\
+    cp LICENSE.md /clair3
+
+RUN mkdir /clair3/models &&\
+    wget http://www.bio8.cs.hku.hk/clair3/clair3_models/clair3_models.tar.gz &&\    
+    tar --no-same-owner -C /clair3/models -xvf clair3_models.tar.gz
+
+## App ##
+FROM mambaorg/micromamba:1.5.8 as app
+
+ARG CLAIR3_VER
+
+USER root
+
+WORKDIR /
+
+LABEL base.image="mambaorg/micromamba:1.5.8"
+LABEL dockerfile.version="1"
+LABEL software="CLAIR3"
+LABEL software.version="${CLAIR3_VER}"
+LABEL description="Clair3 is a germline small variant caller for long-reads."
+LABEL website="https://github.com/HKU-BAL/Clair3"
+LABEL license="https://github.com/HKU-BAL/Clair3/blob/main/LICENSE.md"
+LABEL maintainer="Kutluhan Incekara"
+LABEL maintainer.email="[email protected]"
+
+RUN apt-get update && apt-get install --no-install-recommends -y\
+    procps
+
+RUN micromamba install --name base -c conda-forge -c bioconda \
+    python=3.9.0 \
+    numpy=1.24.3 \
+    pypy3.6 \
+    tensorflow-cpu=2.8.0 \
+    pytables \
+    pigz \
+    cffi=1.14.4 \
+    parallel=20191122 \
+    zstd \
+    samtools=1.15.1 \
+    whatshap=1.7 &&\
+    micromamba clean -a -y &&\
+    rm -rf /opt/conda/pkgs/
+
+ENV PATH="/opt/conda/bin/:/clair3:${PATH}" \
+    LC_ALL=C.UTF-8
+
+RUN pypy3 -m ensurepip &&\
+    pypy3 -m pip install --no-cache mpmath==1.2.1 &&\
+    pip install --no-cache tensorflow-addons
+
+COPY --from=builder /clair3 /clair3
+
+CMD run_clair3.sh
+
+WORKDIR /data
+
+## Test ##
+FROM app as test
+
+RUN apt-get update && apt-get install -y wget
+
+COPY ont_quick_test.sh . 
+
+RUN chmod +x ont_quick_test.sh &&\
+    ./ont_quick_test.sh

clair3/1.0.9/README.md

@@ -0,0 +1,32 @@
+# Clair3 container
+
+Main tool: [clair3](https://github.com/HKU-BAL/Clair3)
+
+Code repository: https://github.com/HKU-BAL/Clair3
+
+Basic information on how to use this tool:
+- executable: run_clair3.sh
+- help: -h, --help
+- version: -v, --version
+- description: <tool does something>
+
+Additional information:
+
+This container includes models in `/clair3/models`
+
+Full documentation: https://github.com/HKU-BAL/Clair3
+
+## Example Usage
+
+```bash
+run_clair3.sh \
+  --bam_fn=${BAM} \
+  --ref_fn=${REF} \
+  --threads=${THREADS} \  		     
+  --platform="ont" \               ## options: {ont,hifi,ilmn}
+  --model_path=${MODEL_PREFIX} \   ## absolute model path prefix
+  --output=${OUTPUT_DIR}           ## absolute output path prefix
+## pileup output file: ${OUTPUT_DIR}/pileup.vcf.gz
+## full-alignment output file: ${OUTPUT_DIR}/full_alignment.vcf.gz
+## Clair3 final output file: ${OUTPUT_DIR}/merge_output.vcf.gz
+```

clair3/1.0.9/ont_quick_test.sh

@@ -0,0 +1,44 @@
+## Adapted from http://www.bio8.cs.hku.hk/clair3/demo/clair3_ont_quick_demo.sh
+# Parameters
+PLATFORM='ont'
+INPUT_DIR="/data/clair3_ont_quickDemo"
+OUTPUT_DIR="${INPUT_DIR}/output"
+THREADS=4
+
+## Create local directory structure
+mkdir -p ${INPUT_DIR}
+mkdir -p ${OUTPUT_DIR}
+
+# Download quick demo data
+#GRCh38_no_alt Reference
+wget -q -P ${INPUT_DIR} http://www.bio8.cs.hku.hk/clair3/demo/quick_demo/ont/GRCh38_no_alt_chr20.fa
+wget -q -P ${INPUT_DIR} http://www.bio8.cs.hku.hk/clair3/demo/quick_demo/ont/GRCh38_no_alt_chr20.fa.fai
+# BAM chr20:100000-300000
+wget -q -P ${INPUT_DIR} http://www.bio8.cs.hku.hk/clair3/demo/quick_demo/ont/HG003_chr20_demo.bam
+wget -q -P ${INPUT_DIR} http://www.bio8.cs.hku.hk/clair3/demo/quick_demo/ont/HG003_chr20_demo.bam.bai
+# GIAB Truth VCF and BED
+wget -q -P ${INPUT_DIR} http://www.bio8.cs.hku.hk/clair3/demo/quick_demo/ont/HG003_GRCh38_chr20_v4.2.1_benchmark.vcf.gz
+wget -q -P ${INPUT_DIR} http://www.bio8.cs.hku.hk/clair3/demo/quick_demo/ont/HG003_GRCh38_chr20_v4.2.1_benchmark.vcf.gz.tbi
+wget -q -P ${INPUT_DIR} http://www.bio8.cs.hku.hk/clair3/demo/quick_demo/ont/HG003_GRCh38_chr20_v4.2.1_benchmark_noinconsistent.bed
+
+REF="GRCh38_no_alt_chr20.fa"
+BAM="HG003_chr20_demo.bam"
+BASELINE_VCF_FILE_PATH="HG003_GRCh38_chr20_v4.2.1_benchmark.vcf.gz"
+BASELINE_BED_FILE_PATH="HG003_GRCh38_chr20_v4.2.1_benchmark_noinconsistent.bed"
+OUTPUT_VCF_FILE_PATH="merge_output.vcf.gz"
+
+CONTIGS="chr20"
+START_POS=100000
+END_POS=300000
+echo -e "${CONTIGS}\t${START_POS}\t${END_POS}" > ${INPUT_DIR}/quick_demo.bed
+
+cd ${OUTPUT_DIR}
+# Run Clair3 using one command
+  run_clair3.sh \
+  --bam_fn=${INPUT_DIR}/${BAM} \
+  --ref_fn=${INPUT_DIR}/${REF} \
+  --threads=${THREADS} \
+  --platform=${PLATFORM} \
+  --model_path="/clair3/models/${PLATFORM}" \
+  --output=${OUTPUT_DIR} \
+  --bed_fn=${INPUT_DIR}/quick_demo.bed