Poor Alignment with Existing Ontologies and Standards

[dragon-ai-agent]

Problem

While BioStride mentions ontology integration, the schema uses mostly custom enums and free-text fields where community ontologies could provide greater semantic precision. There are no explicit mappings to OBI, UBERON, or other standards.

Current State

• Empty mappings: Schema classes show empty "Mappings" sections or just self-references
• Unmapped enumerations: TechniqueEnum, PreparationTypeEnum have no ontology CURIEs
• Missing OBI alignment: ExperimentRun, SamplePreparation could map to OBI assay/process terms
• No technique mappings: SAXS, cryo-EM could reference EDAM or OBI technique terms
• Undefined quality metrics: Resolution, Rg, completeness lack standardized definitions

Examples of missing mappings:

• ExperimentRun → could map to OBI "data acquisition"
• SamplePreparation → could map to OBI sample preparation processes
• Instrument → could map to OBI "instrument device" or BFO "device"
• sample_type enum values → could map to OBI or Sample Ontology terms

Impact

• Poor interoperability: Cannot easily integrate with other ontology-annotated datasets
• Semantic isolation: Creates mini-ontology instead of reusing community standards
• Query limitations: Cannot leverage semantic web tools for BioStride data
• Definition gaps: Some classes lack descriptions that could come from ontologies

Suggested Solutions

1. Add ontology mappings to schema classes

Use LinkML mappings slot:

• ExperimentRun → OBI:0001911 (if relevant for data acquisition)
• SamplePreparation → appropriate OBI process term
• Instrument → BFO "device" or OBI instrument term

2. Embed ontology identifiers in enumerations

Add meaning field to enum values:

• TechniqueEnum.cryo_em → EDAM:operation_364 "cryo EM"
• TechniqueEnum.saxs → EDAM:operation_3450 (or similar)
• Currently these are "None" in docs

3. Expand OntologyTerm usage

Replace hardcoded enums with OntologyTerm references:

• sample_type → OntologyTerm (referencing CLO, Sample Ontology, etc.)
• experiment_type → OntologyTerm for experiment classification
• Follow the existing ImageFeature.terms pattern

4. Link quality metrics to standards

Map QualityMetrics fields (resolution, completeness) to:

• mmCIF/PDBx metadata standards
• Small Angle Scattering community terms
• PDBe metadata definitions

Target Ontologies

• OBI (Ontology for Biomedical Investigations): 2,500+ terms for experiments, assays, devices
• EDAM (bioinformatics operations): Technique and data type terms
• UBERON: Anatomy terms (already mentioned for biological context)
• BFO (Basic Formal Ontology): Upper-level terms for devices, processes
• mmCIF/PDBx: Structural biology metadata standards

Benefits

• Improved interoperability with existing datasets
• Semantic web compatibility for advanced queries
• Reuse of vetted definitions instead of reinventing terms
• Future-proofing through standards alignment

Priority

High - Critical for community adoption and semantic interoperability.