Contact: Ruibang Luo, Zhenxian Zheng
Email: {rbluo,zxzheng}@cs.hku.hk
Clair3-RNA is a small variant caller for RNA long-read data. Clair3-RNA supports ONT complementary DNA sequencing (cDNA) and direct RNA sequencing (dRNA). dRNA sequencing support the ONT latest SQK-RNA004 kit data for variant calling. Clair3-RNA also supports PacBio Sequel and PacBio MAS-Seq RNA sequencing data.
For germline small variant calling, please use Clair3.
For somatic small variant calling using a tumor-normal pair, please try ClairS.
For somatic small variant calling using tumor sample only, please try ClairS-TO.
v0.1.0 (Aug 15, 2024) : 1. Added a new ONT dRNA004 direct RNA sequencing model (ont_dorado_drna004
) for SQK-RNA004 kit. 2. Added new PacBio Sequel (hifi_sequel2_minimap2
) and Revio (hifi_mas_minimap2
) model to support minimap2 alignment. 3. Enhance model training techniques to boost performance by incorporating strategies such as managing low-coverage sites, verifying variant zygosity, filtering RNA editing sites, etc. 4. Renamed all ONT and PacBio model names, check here for more details.
v0.0.1 (Nov 27, 2023): Initial release for early access.
- Oxford Nanopore (ONT) data as input, see ONT Quick Demo.
- PacBio HiFi data as input, see PacBio HiFi Quick Demo.
After following installation, you can run Clair3-RNA with one command:
./run_clair3_rna -B input.bam -R ref.fa -o output -t 8 -p ont_dorado_drna004
## Final output file: output/output.vcf.gz
Check Usage for more options.
Clair-RNA was trained using GIAB RNA sequencing data. All models were trained with chr20 excluded (including only chr1-19, 21, 22).
Platform | Chemistry/Kit/Instruments | Basecaller | Aligner | Option (-p/--platform ) |
Reference | Training samples |
---|---|---|---|---|---|---|
ONT | SQK-RNA004 kit, direct RNA sequencing | Dorado | minimap2 | ont_dorado_drna004 |
GRCh38 | HG002 |
ONT | SQK-RNA002 kit, direct RNA sequencing | Guppy | minimap2 | ont_guppy_drna002 |
GRCh38 | HG002 |
ONT | R9.4.1, complementary DNA sequencing | Guppy | minimap2 | ont_guppy_cdna |
GRCh38 | HG002 |
PacBio HiFi | Sequel with Iso-Seq kit | - | pbmm2/minimap2 | hifi_sequel2_pbmm2 , hifi_sequel2_minimap2 |
GRCh38 | HG002 |
PacBio HiFi | Revio with MAS-Seq kit | - | pbmm2/minimap2 | hifi_mas_pbmm2 , hifi_mas_minimap2 |
GRCh38 | HG002 |
A pre-built docker image is available at DockerHub.
Caution: Absolute path is needed for both INPUT_DIR
and OUTPUT_DIR
in docker.
docker run -it \
-v ${INPUT_DIR}:${INPUT_DIR} \
-v ${OUTPUT_DIR}:${OUTPUT_DIR} \
hkubal/clair3-rna:latest \
/opt/bin/run_clair3_rna \
--bam_fn ${INPUT_DIR}/input.bam \ ## use your input bam file name here
--ref_fn ${INPUT_DIR}/ref.fa \ ## use your reference file name here
--threads ${THREADS} \ ## maximum threads to be used
--platform ${PLATFORM} \ ## options: {ont_dorado_drna004, ont_guppy_drna002, ont_guppy_cdna, hifi_sequel2_pbmm2, hifi_sequel2_minimap2, hifi_mas_pbmm2, hifi_sequel2_minimap2}
--tag_variant_using_readiportal ## optional, tag variants using REDIportal dataset
--output_dir ${OUTPUT_DIR} ## output path prefix
Check Usage for more options.
Caution: Absolute path is needed for both INPUT_DIR
and OUTPUT_DIR
in singularity.
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)
mkdir -p ${OUTPUT_DIR}
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-rna:latest
# run the sandbox like this afterward
singularity exec \
-B ${INPUT_DIR},${OUTPUT_DIR} \
clair3-rna_latest.sif \
/opt/bin/run_clair3_rna \
--bam_fn ${INPUT_DIR}/input.bam \ ## use your input bam file name here
--ref_fn ${INPUT_DIR}/ref.fa \ ## use your reference file name here
--threads ${THREADS} \ ## maximum threads to be used
--platform ${PLATFORM} \ ## options: {ont_dorado_drna004, ont_guppy_drna002, ont_guppy_cdna, hifi_sequel2_pbmm2, hifi_sequel2_minimap2, hifi_mas_pbmm2, hifi_sequel2_minimap2}
--tag_variant_using_readiportal ## optional, tag variants using REDIportal dataset
--output_dir ${OUTPUT_DIR} \ ## output path prefix
--conda_prefix /opt/conda/envs/clair3_rna
Anaconda install:
Please install anaconda using the official guide or using the commands below:
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
Install Clair3-RNA using anaconda step by step:
# create and activate an environment named clair3_rna
# install pypy and packages in the environment
conda create -n clair3_rna -c conda-forge -c bioconda clair3 mosdepth bedtools -y
source activate clair3_rna
git clone https://github.com/HKU-BAL/Clair3-RNA.git
cd Clair-RNA
# make sure in conda environment
# download pre-trained models
echo ${CONDA_PREFIX}
mkdir -p ${CONDA_PREFIX}/bin/clair3_rna_models
wget http://www.bio8.cs.hku.hk/clair3_rna/models/clair3_rna_models.tar.gz
tar -zxvf clair3_rna_models.tar.gz -C ${CONDA_PREFIX}/bin/clair3_rna_models/
./run_clair3_rna --help
This is the same as option 1 except that you are building a docker image yourself. Please refer to option 1 for usage.
git clone https://github.com/HKU-BAL/Clair3-RNA.git
cd Clair-RNA
# build a docker image named hkubal/clairs:latest
# might require docker authentication to build docker image
docker build -f ./Dockerfile -t hkubal/clair3-rna:latest .
# run the docker image like option 1
docker run -it hkubal/clair3-rna:latest /opt/bin/clair3_rna --help
./run_clair3_rna \
--bam_fn ${INPUT_DIR}/input.bam \ ## use your input bam file name here
--ref_fn ${INPUT_DIR}/ref.fa \ ## use your reference file name here
--threads ${THREADS} \ ## maximum threads to be used
--platform ${PLATFORM} \ ## options: {ont_dorado_drna004, ont_guppy_drna002, ont_guppy_cdna, hifi_sequel2_pbmm2, hifi_sequel2_minimap2, hifi_mas_pbmm2, hifi_sequel2_minimap2}
--tag_variant_using_readiportal ## optional, tag variants using REDIportal dataset
--output_dir ${OUTPUT_DIR} ## output path prefix
## Final output file: ${OUTPUT_DIR}/output.vcf.gz
Required parameters:
-B BAM_FN, --bam_fn BAM_FN
RNA BAM file input. The input file must be samtools indexed.
-R REF_FN, --ref_fn REF_FN
FASTA reference file input. The input file must be samtools indexed.
-o OUTPUT_DIR, --output_dir OUTPUT_DIR
VCF output directory.
-t THREADS, --threads THREADS
Max #threads to be used.
-p PLATFORM, --platform PLATFORM
Select the sequencing platform of the input. Possible options: {ont_dorado_drna004, ont_guppy_drna002, ont_guppy_cdna, hifi_sequel2_pbmm2, hifi_sequel2_minimap2, hifi_mas_pbmm2, hifi_sequel2_minimap2}.
Miscellaneous parameters:
-P PILEUP_MODEL_PATH, --pileup_model_path PILEUP_MODEL_PATH
Specify the path prefix to your own pileup model. Including ${pileup_model_path}.data-00000-of-00001, ${pileup_model_path}.index.
-c CTG_NAME, --ctg_name CTG_NAME
The name of the contigs to be processed. Split by ',' for multiple contigs. Default: call in chr{1..22} and {1..22}.
-r REGION, --region REGION
A region to be processed. Format: `ctg_name:start-end` (start is 1-based).
-b BED_FN, --bed_fn BED_FN
Path to a BED file. Call variants only in the provided BED regions.
-G GENOTYPING_MODE_VCF_FN, --genotyping_mode_vcf_fn GENOTYPING_MODE_VCF_FN
VCF file input containing candidate sites to be genotyped. Variants will only be called at the sites in the VCF file if provided.
-q QUAL, --qual QUAL If set, variants with >QUAL will be marked as PASS, or LowQual otherwise.
--snp_min_af SNP_MIN_AF
Minimal SNP AF required for a variant to be called. Decrease SNP_MIN_AF might increase a bit of sensitivity, but in trade of precision, speed and accuracy. Default: 0.08.
--indel_min_af INDEL_MIN_AF
Minimum Indel AF required for a candidate variant to be called. Default: ont:0.15,hifi:0.08.
--min_coverage MIN_COVERAGE
Minimal coverage required for a variant to be called. Default: 4.
--tag_variant_using_readiportal
Tag variants uisng REDIportal dataset, If set, called variants that are also in the readiportal dataset will be marked as "RNAEditing". Default: disable.
--readiportal_database_filter_tag READIPORTAL_DATABASE_FILTER_TAG
Use only editing sites with these tags in the readiportal dataset, split by ":" for multiple tags. Default: using sites supported by two or more sources - "A,D:A,R:A,R,D".
--readiportal_reference_genome_version READIPORTAL_REFERENCE_GENOME_VERSION
Select the reference genome version in the readiportal dataset. Possible options: {grch38, grch37}. Default: "grch38".
--chunk_size CHUNK_SIZE
The size of each chuck for parallel processing. Default: 5000000.
-s SAMPLE_NAME, --sample_name SAMPLE_NAME
Define the sample name to be shown in the VCF file. Default: SAMPLE.
--output_prefix OUTPUT_PREFIX
Prefix for output VCF filename. Default: output.
--remove_intermediate_dir
Remove intermediate directory before finishing to save disk space.
--include_all_ctgs Call variants on all contigs, otherwise call in chr{1..22} and {1..22}.
--print_ref_calls Show reference calls (0/0) in VCF file.
-d, --dry_run Print the commands that will be ran.
--python PYTHON Absolute path of python, python3 >= 3.9 is required.
--pypy PYPY Absolute path of pypy3, pypy3 >= 3.6 is required.
--samtools SAMTOOLS Absolute path of samtools, samtools version >= 1.10 is required.
--parallel PARALLEL Absolute path of parallel, parallel >= 20191122 is required.
--min_mq MIN_MQ Minimal mapping quality required for an alignment to be considered. Default: 5.
./run_clair3_rna -B input.bam -R ref.fa -o output -t 8 -p ont_dorado_drna004 -C chr21,chr22
./run_clair3_rna -B input.bam -R ref.fa -o output -t 8 -p ont_dorado_drna004 -r chr20:1000000-2000000
Call variants at interested variant sites (genotyping) using the -G/--genotyping_mode_vcf_fn
parameter
./run_clair3_rna -B input.bam -R ref.fa -o output -t 8 -p ont_dorado_drna004 -G input.vcf
We highly recommended using BED file to define multiple regions of interest like:
./run_clair3_rna -B input.bam -R ref.fa -o output -t 8 -p ont_dorado_drna004 -b input.bed
RNA undergoes editing by ADAR (adenosine deaminases acting on RNA), resulting in Adenosine-to-inosine (A-to-I) changes. These A-to-I changes can be observed in RNA-seq datasets as A-to-G and T-to-C changes, which do not represent genuine RNA variants. To address this, we provide users with the option to utilize external datasets such as REDIportal to annotate RNA editing sites. In Clair3-RNA's VCF output, variants that are also RNA editing sites reported in REDIportal can be tagged. These sites will be marked as RNAEditing
instead of PASS
in the FILTER
column when the --tag_variant_using_readiportal
option is enabled.
Caution: --tag_variant_using_readiportal
option currently works for GRCh38 and GRCh37 reference genome only, use can specify the reference genome version by using option --readiportal_reference_genome_version={grch38, grch37}
.