CNAG Biomedical Informatics framework for variant calling
CBIcall (CNAG Biomedical Informatics framework for variant calling) is a configuration-driven framework for reproducible variant calling in large sequencing cohorts. It validates analysis intent, resolves approved workflow backends and resource definitions, and records structured run reports for auditable production runs and run-to-run comparison.
Documentation: https://cnag-biomedical-informatics.github.io/cbicall/
CBIcall orchestrates germline variant calling workflows for Illumina sequencing data. It does not implement variant calling algorithms itself. Instead, it validates parameters, resolves workflows from a versioned registry, checks resource compatibility, launches native CBIcall workflows through Bash, Snakemake, Nextflow, or Cromwell backends, and captures audit artifacts for traceability. Selected external nf-core workflows can also run through the same validation and provenance layer.
CBIcall uses a three-part execution contract:
| Contract layer | Purpose |
|---|---|
| Parameters YAML | User analysis intent: inputs, pipeline, mode, genome, backend, and runtime options. |
| Workflow registry | Approved workflow implementations and backend-specific entrypoints. |
| Resource catalog | External references, tool/resource bundles, compatibility rules, and resource identity. |
Key points:
- Configuration-driven execution from a YAML parameter file
- Native CBIcall workflow support through Bash, Snakemake, Nextflow, and Cromwell backends
- Support for WES, WGS, and mtDNA analysis modes
- Registry-backed support for selected external nf-core/Nextflow workflows
- Structured audit artifacts:
log.json,run-report.json, optionalrun-report.html, workflow fingerprints, resource identity, output inventories, and normalized VCF hashes - Programmatic run comparison with
cbicall compare-runs - Optional partial workflow starts for supported backends
Workflow sources:
| Source | Role |
|---|---|
| Native CBIcall workflows | Packaged WES/WGS/mtDNA pipelines with CBIcall validation, logging, and output structure. |
| External nf-core workflows | Selected registry-backed Nextflow workflows executed with CBIcall validation and provenance. |
Install the core command from PyPI:
python3 -m pip install cbicallFor development or all optional Python integrations:
python3 -m pip install -e ".[all,test]" # source checkout
python3 -m pip install "cbicall[all]" # released packageNative workflows also require the separately distributed resource bundle. See the documentation for Python, Docker, Apptainer, cloud, and HPC installation instructions.
Point all native backends to an installed bundle with:
export CBICALL_DATA=/absolute/path/to/cbicall-data
cbicall validate-resourcesGenerate a self-contained tour of the WES audit report and interactive mtDNA browser without installing the external resource bundle:
cbicall demoThis uses precomputed outputs from the packaged integration fixture. It does not execute BWA, GATK, or MToolBox. For a real analysis, use a parameters YAML:
cbicall run -p params.yaml -t 8Runnable examples and sample inputs are available under examples/.
CBIcall: a configuration-driven framework for variant calling in large sequencing cohorts. Preprint DOI.
Written by Manuel Rueda (mrueda). GitHub repository: https://github.com/CNAG-Biomedical-Informatics/cbicall.
Please see the GPLv3 license for distribution and usage terms.