The Bento Platform started as part of the EpiShare GA4GH Driver Project. Bento puts GA4GH standards, schemas and formats as part of its core, and extends them when they’re not sufficient. Any concept that has a corresponding GA4GH standard should be implemented in Bento following that standard.
- Clinical / Phenotypic data -> Phenopackets
- Variants -> VCF and VRS
- Obtaining a reference genome -> refget
- Discovery -> Beacon v2 API
- File access API -> Data Repository Service
- Workflow execution -> WES
Different Bento setups could make use of some services, but not others. For instance, one Bento deployment could choose to offer the cBioPortal as an additional way to query data, but another project might not have a need for it.
- Data types have their own data service
- APIs are also separated in services
- Data Search using the Beacon API
- Data ingest using WES-formatted workflows
- They all implement GA4GH Service Registry
- All services are implemented as Docker containers
- A Bento deployment is a docker-compose setup deployed on a server
- The platform is implemented for multiple projects, supporting multiple organisms
- Goal: Develop things only once for all supported projects and avoid maintaining multiple forks for different portals
- Adapt a project (meta)data to commonly used standards, increasing compatibility and discoverability with datasets from other genomics initiatives
- Customisation of an instance is done through parameters, config files, images, etc.
- Deployment of a Bento instance does not require advanced knowledge of the stack
- CLI tooling is enough to launch a new instance