cd ${HOME}
wget "http://www.bio8.cs.hku.hk/clair3/demo/clair3_ont_quick_demo.sh"
chmod +x clair3_ont_quick_demo.sh
./clair3_ont_quick_demo.sh

INPUT_DIR="[YOUR_INPUT_FOLDER]"        # e.g. /home/user1/input (absolute path needed)
OUTPUT_DIR="[YOUR_OUTPUT_FOLDER]"      # e.g. /home/user1/output (absolute path needed)
THREADS="[MAXIMUM_THREADS]"            # e.g. 8
MODEL_NAME="[YOUR_MODEL_NAME]"         # e.g. r941_prom_hac_g360+g422

docker run -it \
  -v ${INPUT_DIR}:${INPUT_DIR} \
  -v ${OUTPUT_DIR}:${OUTPUT_DIR} \
  hkubal/clair3:latest \
  /opt/bin/run_clair3.sh \
  --bam_fn=${INPUT_DIR}/input.bam \    ## change your bam file name here
  --ref_fn=${INPUT_DIR}/ref.fa \       ## change your reference file name here
  --threads=${THREADS} \               ## maximum threads to be used
  --platform="ont" \                   ## options: {ont,hifi,ilmn}
  --model_path="/opt/models/${MODEL_NAME}" \
  --output=${OUTPUT_DIR}               ## absolute output path prefix 

INPUT_DIR="[YOUR_INPUT_FOLDER]"        # e.g. /home/user1/input (absolute path needed)
OUTPUT_DIR="[YOUR_OUTPUT_FOLDER]"      # e.g. /home/user1/output (absolute path needed)
THREADS="[MAXIMUM_THREADS]"            # e.g. 8
MODEL_NAME="[YOUR_MODEL_NAME]"         # e.g. r941_prom_hac_g360+g422

conda config --add channels defaults
conda create -n singularity-env -c conda-forge singularity -y
conda activate singularity-env

# singularity pull docker pre-built image
singularity pull docker://hkubal/clair3:latest

# run clair3 like this afterward
singularity exec clair3_latest.sif \
  /opt/bin/run_clair3.sh \
  --bam_fn=${INPUT_DIR}/input.bam \    ## change your bam file name here
  --ref_fn=${INPUT_DIR}/ref.fa \       ## change your reference file name here
  --threads=${THREADS} \               ## maximum threads to be used
  --platform="ont" \                   ## options: {ont,hifi,ilmn}
  --model_path="/opt/models/${MODEL_NAME}" \
  --output=${OUTPUT_DIR}               ## absolute output path prefix

# make sure channels are added in conda
conda config --add channels defaults
conda config --add channels bioconda
conda config --add channels conda-forge

# create conda environment named "clair3"
# replace clair3 by clair3-illumina for using illumina data
conda create -n clair3 -c bioconda clair3 python=3.9.0 -y
conda activate clair3

# run clair3 like this afterward
MODEL_NAME="[YOUR_MODEL_NAME]"         # e.g. r941_prom_hac_g360+g422

run_clair3.sh \
  --bam_fn=input.bam \                 ## change your bam file name here
  --ref_fn=ref.fa \                    ## change your reference file name here
  --threads=${THREADS} \               ## maximum threads to be used
  --platform="ont" \                   ## options: {ont,hifi,ilmn}
  --model_path="${CONDA_PREFIX}/bin/models/${MODEL_NAME}" \ 
  --output=${OUTPUT_DIR}               ## output path prefix 

wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
chmod +x ./Miniconda3-latest-Linux-x86_64.sh 
./Miniconda3-latest-Linux-x86_64.sh

INPUT_DIR="[YOUR_INPUT_FOLDER]"        # e.g. ./input
OUTPUT_DIR="[YOUR_OUTPUT_FOLDER]"      # e.g. ./output
THREADS="[MAXIMUM_THREADS]"            # e.g. 8

# create and activate an environment named clair3
conda create -n clair3 python=3.9.0 -y
source activate clair3

# install pypy and packages in the environemnt
conda install -c conda-forge pypy3.6 -y
pypy3 -m ensurepip
pypy3 -m pip install mpmath==1.2.1

# install python packages in environment
conda install -c conda-forge tensorflow==2.8.0 -y
conda install -c conda-forge pytables -y
conda install -c anaconda pigz cffi==1.14.4 -y
conda install -c conda-forge parallel=20191122 zstd -y
conda install -c conda-forge -c bioconda samtools=1.15.1 -y
conda install -c conda-forge -c bioconda whatshap=1.4 -y
conda install -c conda-forge xz zlib bzip2 automake curl -y
# tensorflow-addons is required in training
pip install tensorflow-addons

# clone Clair3
git clone https://github.com/HKU-BAL/Clair3.git
cd Clair3

# compile samtools, longphase and cffi library for c implement
# after building, longphase binary is in `Clair3` folder
source activate clair3 && make PREFIX=${CONDA_PREFIX}

# download pre-trained models
mkdir models
wget http://www.bio8.cs.hku.hk/clair3/clair3_models/clair3_models.tar.gz 
tar -zxvf clair3_models.tar.gz -C ./models

# run clair3
MODEL_NAME="[YOUR_MODEL_NAME]"         # e.g. r941_prom_hac_g360+g422
./run_clair3.sh \
  --bam_fn=${INPUT_DIR}/input.bam \    ## change your bam file name here
  --ref_fn=${INPUT_DIR}/ref.fa \       ## change your reference file name here
  --threads=${THREADS} \               ## maximum threads to be used
  --platform="ont" \                   ## options: {ont,hifi,ilmn}
  --model_path=`pwd`"/models/${MODEL_NAME}" \
  --output=${OUTPUT_DIR}               ## output path prefix

# clone Clair3
git clone https://github.com/hku-bal/Clair3.git
cd Clair3

# build a docker image named hkubal/clair3:latest
# might require docker authentication to build docker image 
docker build -f ./Dockerfile -t hkubal/clair3:latest .

# run clair3 docker image like option 1
docker run -it hkubal/clair3:latest /opt/bin/run_clair3.sh --help

./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

  -b, --bam_fn=FILE             BAM file input. The input file must be samtools indexed.
  -f, --ref_fn=FILE             FASTA reference file input. The input file must be samtools indexed.
  -m, --model_path=STR          The folder path containing a Clair3 model (requiring six files in the folder, including pileup.data-00000-of-00002, pileup.data-00001-of-00002 pileup.index, full_alignment.data-00000-of-00002, full_alignment.data-00001-of-00002  and full_alignment.index).
  -t, --threads=INT             Max threads to be used. The full genome will be divided into small chunks for parallel processing. Each chunk will use 4 threads. The chunks being processed simultaneously is ceil($threads/4)*3. 3 is the overloading factor.
  -p, --platform=STR            Select the sequencing platform of the input. Possible options: {ont,hifi,ilmn}.
  -o, --output=PATH             VCF/GVCF output directory.

      --bed_fn=FILE             Call variants only in the provided bed regions.
      --vcf_fn=FILE             Candidate sites VCF file input, variants will only be called at the sites in the VCF file if provided.
      --ctg_name=STR            The name of the sequence to be processed.
      --sample_name=STR         Define the sample name to be shown in the VCF file.
      --qual=INT                If set, variants with >$qual will be marked PASS, or LowQual otherwise.
      --samtools=STR            Path of samtools, samtools version >= 1.10 is required.
      --python=STR              Path of python, python3 >= 3.6 is required.
      --pypy=STR                Path of pypy3, pypy3 >= 3.6 is required.
      --parallel=STR            Path of parallel, parallel >= 20191122 is required.
      --whatshap=STR            Path of whatshap, whatshap >= 1.0 is required.
      --longphase=STR           Path of longphase, longphase >= 1.0 is required.
      --chunk_size=INT          The size of each chuck for parallel processing, default: 5Mbp.
      --pileup_only             Use the pileup model only when calling, default: disable.
      --print_ref_calls         Show reference calls (0/0) in vcf file, default: disable.
      --include_all_ctgs        Call variants on all contigs, otherwise call in chr{1..22,X,Y} and {1..22,X,Y}, default: disable.
      --gvcf                    Enable GVCF output, default: disable.
      --enable_phasing          Output phased variants using whatshap, default: disable.
      --longphase_for_phasing   Use longphase for phasing, default: enable.
      --disable_c_impl          Disable C implement with cffi for pileup and full-alignment create tensor, default: enable.
      --remove_intermediate_dir Remove intermediate directory, including intermediate phased BAM, pileup and full-alignment results. default: disable.
      --snp_min_af=FLOAT        Minimum SNP AF required for a candidate variant. Lowering the value might increase a bit of sensitivity in trade of speed and accuracy, default: ont:0.08,hifi:0.08,ilmn:0.08.
      --indel_min_af=FLOAT      Minimum INDEL AF required for a candidate variant. Lowering the value might increase a bit of sensitivity in trade of speed and accuracy, default: ont:0.15,hifi:0.08,ilmn:0.08.
      --var_pct_full=FLOAT      EXPERIMENTAL: Specify an expected percentage of low quality 0/1 and 1/1 variants called in the pileup mode for full-alignment mode calling, default: 0.3.
      --ref_pct_full=FLOAT      EXPERIMENTAL: Specify an expected percentage of low quality 0/0 variants called in the pileup mode for full-alignment mode calling, default: 0.3 for ilmn and hifi, 0.1 for ont.
      --var_pct_phasing=FLOAT   EXPERIMENTAL: Specify an expected percentage of high quality 0/1 variants used in WhatsHap phasing, default: 0.8 for ont guppy5 and 0.7 for other platforms.
      --pileup_model_prefix=STR EXPERIMENTAL: Model prefix in pileup calling, including $prefix.data-00000-of-00002, $prefix.data-00001-of-00002 $prefix.index. default: pileup.
      --fa_model_prefix=STR     EXPERIMENTAL: Model prefix in full-alignment calling, including $prefix.data-00000-of-00002, $prefix.data-00001-of-00002 $prefix.index, default: full_alignment.
      --min_mq=INT              EXPERIMENTAL: If set, reads with mapping quality with <$min_mq are filtered, default: 5.
      --min_coverage=INT        EXPERIMENTAL: Minimum coverage required to call a variant, default: 2.
      --min_contig_size=INT     EXPERIMENTAL: If set, contigs with contig size<$min_contig_size are filtered, default: 0.
      --fast_mode               EXPERIMENTAL: Skip variant candidates with AF <= 0.15, default: disable.
      --haploid_precise         EXPERIMENTAL: Enable haploid calling mode. Only 1/1 is considered as a variant, default: disable.
      --haploid_sensitive       EXPERIMENTAL: Enable haploid calling mode. 0/1 and 1/1 are considered as a variant, default: disable.
      --no_phasing_for_fa       EXPERIMENTAL: Call variants without whatshap phasing in full alignment calling, default: disable.
      --call_snp_only           EXPERIMENTAL: Call candidates pass SNP minimum AF only, ignore Indel candidates, default: disable.
      --enable_long_indel       EXPERIMENTAL: Call long Indel variants(>50 bp), default: disable.

CONTIGS_LIST="[YOUR_CONTIGS_LIST]"     # e.g "chr21" or "chr21,chr22"
INPUT_DIR="[YOUR_INPUT_FOLDER]"        # e.g. /home/user1/input  (absolute path needed)
OUTPUT_DIR="[YOUR_OUTPUT_FOLDER]"      # e.g. /home/user1/output (absolute path needed)
THREADS="[MAXIMUM_THREADS]"            # e.g. 8
MODEL_NAME="[YOUR_MODEL_NAME]"         # e.g. r941_prom_hac_g360+g422

docker run -it \
  -v ${INPUT_DIR}:${INPUT_DIR} \
  -v ${OUTPUT_DIR}:${OUTPUT_DIR} \
  hkubal/clair3:latest \
  /opt/bin/run_clair3.sh \
  --bam_fn=${INPUT_DIR}/input.bam \    ## change your bam file name here
  --ref_fn=${INPUT_DIR}/ref.fa \       ## change your reference file name here
  --threads=${THREADS} \               ## maximum threads to be used
  --platform="ont" \                   ## options: {ont,hifi,ilmn}
  --model_path="/opt/models/${MODEL_NAME}" \
  --output=${OUTPUT_DIR} \             ## absolute output path prefix
  --ctg_name=${CONTIGS_LIST}

KNOWN_VARIANTS_VCF="[YOUR_VCF_PATH]"   # e.g. /home/user1/known_variants.vcf.gz (absolute path needed)
INPUT_DIR="[YOUR_INPUT_FOLDER]"        # e.g. /home/user1/input (absolute path needed)
OUTPUT_DIR="[YOUR_OUTPUT_FOLDER]"      # e.g. /home/user1/output (absolute path needed)
THREADS="[MAXIMUM_THREADS]"            # e.g. 8
MODEL_NAME="[YOUR_MODEL_NAME]"         # e.g. r941_prom_hac_g360+g422

docker run -it \
  -v ${INPUT_DIR}:${INPUT_DIR} \
  -v ${OUTPUT_DIR}:${OUTPUT_DIR} \
  hkubal/clair3:latest \
  /opt/bin/run_clair3.sh \
  --bam_fn=${INPUT_DIR}/input.bam \    ## change your bam file name here
  --ref_fn=${INPUT_DIR}/ref.fa \       ## change your reference file name here
  --threads=${THREADS} \               ## maximum threads to be used
  --platform="ont" \                   ## options: {ont,hifi,ilmn}
  --model_path="/opt/models/${MODEL_NAME}" \
  --output=${OUTPUT_DIR} \             ## absolute output path prefix
  --vcf_fn=${KNOWN_VARIANTS_VCF}

# define 0-based "ctg start end" if at specific sites
CONTIGS="[YOUR_CONTIGS_NAME]"          # e.g. chr22
START_POS="[YOUR_START_POS]"           # e.g. 0
END_POS="[YOUR_END_POS]"               # e.g 10000
echo -e "${CONTIGS}\t${START_POS}\t${END_POS}" > /home/user1/tmp.bed ## change directory accordingly

BED_FILE_PATH="[YOUR_BED_FILE]"        # e.g. /home/user1/tmp.bed (absolute path needed)
INPUT_DIR="[YOUR_INPUT_FOLDER]"        # e.g. /home/user1/input (absolute path needed)
OUTPUT_DIR="[YOUR_OUTPUT_FOLDER]"      # e.g. /home/user1/output (absolute path needed)
THREADS="[MAXIMUM_THREADS]"            # e.g. 8
MODEL_NAME="[YOUR_MODEL_NAME]"         # e.g. r941_prom_hac_g360+g422

docker run -it \
  -v ${INPUT_DIR}:${INPUT_DIR} \
  -v ${OUTPUT_DIR}:${OUTPUT_DIR} \
  hkubal/clair3:latest \
  /opt/bin/run_clair3.sh \
  --bam_fn=${INPUT_DIR}/input.bam \    ## change your bam file name here
  --ref_fn=${INPUT_DIR}/ref.fa \       ## change your reference file name here
  --threads=${THREADS} \               ## maximum threads to be used
  --platform="ont" \                   ## options: {ont,hifi,ilmn}
  --model_path="/opt/models/${MODEL_NAME}" \
  --output=${OUTPUT_DIR} \             ## absolute output path prefix
  --bed_fn=${BED_FILE_PATH}

INPUT_DIR="[YOUR_INPUT_FOLDER]"        # e.g. /home/user1/input (absolute path needed)
OUTPUT_DIR="[YOUR_OUTPUT_FOLDER]"      # e.g. /home/user1/output (absolute path needed)
THREADS="[MAXIMUM_THREADS]"            # e.g. 8
MODEL_NAME="[YOUR_MODEL_NAME]"         # e.g. r941_prom_hac_g360+g422

docker run -it \
  -v ${INPUT_DIR}:${INPUT_DIR} \
  -v ${OUTPUT_DIR}:${OUTPUT_DIR} \
  hkubal/clair3:latest \
  /opt/bin/run_clair3.sh \
  --bam_fn=${INPUT_DIR}/input.bam \    ## change your bam file name here
  --ref_fn=${INPUT_DIR}/ref.fa \       ## change your reference file name here
  --threads=${THREADS} \               ## maximum threads to be used
  --platform="ont" \                   ## options: {ont,hifi,ilmn}
  --model_path="/opt/models/${MODEL_NAME}" \
  --output=${OUTPUT_DIR} \
  --no_phasing_for_fa \                ## disable phasing for full-alignment
  --include_all_ctgs \                 ## call variants on all contigs in the reference fasta
  --haploid_precise                    ## optional(enable --haploid_precise or --haploid_sensitive) for haploid calling