Owner: Shahin Mohammadi · Created: 2026-05-24 · Status: DRAFT
Canonical location: ~/repos/cytognosis/org/plans/research/biotools-schema-edam-research.md
| # | Section | Purpose |
|---|---------|---------|
| 1 | Executive Summary | Key findings and relevance to Cytognosis |
| 2 | biotoolsSchema Specification | Full schema structure and controlled vocabularies |
| 3 | EDAM Ontology | Four branches, hierarchy, and classification system |
| 4 | bio.tools Registry | Registration, API, curation model, and statistics |
| 5 | Ecosystem Integration | How these components connect to workflow systems |
| 6 | Integration with Cytognosis | Mapping to our schemas, registries, and platform |
| 7 | Recommendations | Concrete next steps |
The biotoolsSchema, EDAM ontology, and bio.tools registry form a tightly integrated "stack" for describing, classifying, and discovering bioinformatics software. Together they represent the ELIXIR research infrastructure's answer to FAIR software principles.
1. biotoolsSchema defines 50+ attributes organized into 9 logical groupings, with only 3 mandatory fields (name, description, homepage). It provides 17 internal controlled vocabularies for technical metadata and relies on EDAM for scientific classification.
2. EDAM (EMBRACE Data and Methods) is a domain ontology with ~3,400+ concepts organized into four branches: Topic, Operation, Data (including Identifiers), and Format. Each forms a directed acyclic graph (DAG) of is_a relationships.
3. bio.tools hosts 33,000+ tool entries with 564,000+ metadata annotations and 211,000+ EDAM annotations from 15,595 contributors. It provides a REST API for programmatic search and registration.
4. Relevance to Cytognosis: These standards directly inform our LEGO biological model registry design, skill-tagging ontology, and experiment orchestration system. EDAM concepts map naturally to our existing LinkML schemas and topic hierarchy.
biotoolsSchema is a formalized description model for bioinformatics software, available as both XML Schema (XSD) and JSON Schema. It defines the syntax, structure, and semantics for cataloging computational tools in the bio.tools registry.
The schema is part of a broader technology stack:
| Component | Role |
|-----------|------|
| biotoolsCore | Flat list of 50+ attributes |
| biotoolsSchema | Formal technical implementation (XSD/JSON Schema) |
| EDAM Ontology | Semantic layer for standardized terminology |
| Curation Guidelines | Human-readable conventions for quality/consistency |
The ~50 attributes are organized into 9 logical groupings:
| # | Group | Key Attributes | Purpose |
|---|-------|---------------|---------|
| 1 | Summary | name, description, homepage, biotoolsID, biotoolsCURIE, version, otherID | Basic identification |
| 2 | Function | operation, input, output (each with EDAM data + format) | What the tool does |
| 3 | Labels | topic, toolType, operatingSystem, language, license, cost, maturity, accessibility | Classification metadata |
| 4 | Links | issue tracker, repository, social media, mailing list | External resources |
| 5 | Download | source code, container image, VM image, Galaxy wrapper | Acquisition methods |
| 6 | Documentation | API docs, user manual, training material | Learning resources |
| 7 | Publication | primary publication, review, method, usage | Literature citations |
| 8 | Relation | isNewVersionOf, hasNewVersion, uses, usedBy | Tool relationships |
| 9 | Credit | developer, maintainer, provider, funder, contributor | Attribution |
Only three attributes are mandatory for registration:
This low barrier enables rapid registration while encouraging progressive enrichment.
The Function group is the core of biotoolsSchema. Software is described as having one or more "functions," each containing:
``
Function
├── Operation (1..n) → EDAM Operation terms
│ e.g., "Sequence alignment" (operation_0292)
├── Input (0..n)
│ ├── Data → EDAM Data type
│ │ e.g., "Sequence" (data_2044)
│ └── Format → EDAM Format type
│ e.g., "FASTA" (format_1929)
└── Output (0..n)
├── Data → EDAM Data type
│ e.g., "Sequence alignment" (data_0863)
└── Format → EDAM Format type
e.g., "ClustalW format" (format_1982)
| Field | Cardinality | Requirement | Source |
|-------|-------------|-------------|--------|
| Function | 1..n | Required | Schema |
| Operation | 1..n per function | Required | EDAM |
| Input | 0..n per function | Optional | EDAM |
| Output | 0..n per function | Optional | EDAM |
| Topic | 0..n per tool | Optional | EDAM |
biotoolsSchema defines 17 internal controlled vocabularies for technical aspects:
| Vocabulary | Values (examples) | Purpose |
|-----------|-------------------|---------|
| Tool type | Command-line tool, Web application, Database portal, Web API, Workflow, Suite, Library, Plugin, Script, Desktop application, Bioinformatics portal, Ontology, SPARQL endpoint, Workbench, Web service | Software classification |
| Operating system | Linux, Windows, macOS | Platform support |
| Programming language | C, Python, Java, R, Perl, etc. | Implementation language |
| License | SPDX identifiers (GPL-3.0, MIT, Apache-2.0) | Usage terms |
| Maturity | Mature, Emerging, Deprecated | Development stage |
| Cost | Free of charge, Commercial | Monetary cost |
| Accessibility | Open access, Open access (with restrictions), Restricted access | Non-monetary access barriers |
| Vocabulary | Purpose |
|-----------|---------|
| Link type | Helpdesk, Repository, Social media, Issue tracker |
| Download type | Source code, VM image, Galaxy wrapper, Container |
| Documentation type | API documentation, Manual, Training material |
| Publication type | Primary publication, Review, Method, Usage |
| Relation type | isNewVersionOf, hasNewVersion, uses, usedBy |
| Entity type | Person, Project, Institute, Division, Funding agency |
| Entity role | Developer, Maintainer, Provider, Documentor, Contributor |
biotoolsSchema focuses on metadata for discovery and description, not execution-layer information. It describes what a tool does, not how to invoke it. This makes it complementary to:
EDAM (EMBRACE Data and Methods) is a community-developed domain ontology for bioinformatics, computational biology, and bioimage informatics. It provides standardized terminology for the semantic annotation of tools, workflows, datasets, and services.
Statistics:Broad domains or fields of interest, study, or technology.
| Level | Example Terms |
|-------|--------------|
| Top-level | Biology, Computer science |
| Mid-level | Genomics, Proteomics, Structural biology |
| Specific | Sequence analysis, Molecular genetics, Image analysis |
Functions or processes performed by a tool. Describes what a tool does, not how or in what context.
| Level | Example Terms |
|-------|--------------|
| Top-level | Analysis, Prediction |
| Mid-level | Sequence analysis, Structure analysis |
| Specific | Sequence alignment, Variant calling, Genome indexing |
| Leaf | Pairwise sequence alignment, Constrained sequence alignment |
Analysis
└── Sequence analysis
└── Sequence comparison
└── Sequence alignment
├── Pairwise sequence alignment
├── Multiple sequence alignment
└── Constrained sequence alignment
`
Data Branch
Types of information consumed or produced by tools. Focuses on the semantic content of data rather than its technical representation.
| Level | Example Terms |
|-------|--------------|
| Top-level | Data |
| Mid-level | Sequence, Structure |
| Specific | Sequence record, Sequence alignment, Phylogenetic tree |
| Identifiers | UniProt accession, EC number, Gene symbol |
The Identifier sub-branch is rooted under Data and catalogs types of identifiers used in bioinformatics. Identifiers often include metadata like regular expressions to validate their syntax.
Format Branch
Specific layouts or structures for representing data in files, strings, or messages.
| Level | Example Terms |
|-------|--------------|
| Top-level | Format |
| Mid-level | Textual format, Binary format, XML |
| Specific | FASTA format, PDB format, FASTQ, VCF, BAM |
Cross-branch relationships: Format concepts are formally related to Data concepts via is_format_of relations (e.g., FASTA format is_format_of Sequence).
3.3 Concept Identifiers
EDAM concepts use persistent URIs:
`
Pattern: http://edamontology.org/_
Example: http://edamontology.org/data_0849 → "Sequence record"
Example: http://edamontology.org/operation_0292 → "Sequence alignment"
Example: http://edamontology.org/format_1929 → "FASTA"
Example: http://edamontology.org/topic_3510 → "Proteomics"
`
3.4 Relationship Types
Primary relationship:
is_a (specialization/generalization), forming DAG hierarchies.
Cross-branch relations include:
has_input / has_output (Operation → Data)
is_format_of (Format → Data)
has_topic (Operation → Topic)
3.5 Comparison with Other Classification Systems
| Feature | EDAM Ontology | MeSH | OpenAlex Concepts |
|---------|--------------|------|-------------------|
| Primary Scope | Bioinformatics (tools, data, operations) | Biomedical & Life Sciences | All Scientific Disciplines |
| Curation | Expert-curated / Community | Expert-curated (NLM) | Automated (ML / Clustering) |
| Structure | Hierarchical (DAG) | Poly-hierarchical | Hierarchical (5 levels) |
| Main Use Case | Annotating tools, workflows, & data | Indexing literature (PubMed) | Bibliometric & landscape analysis |
| Term Count | ~3,472 | ~30,000+ | ~65,000+ |
| Coverage | Deep in bioinformatics | Broad in biomedicine | All sciences |
Mapping efforts: Projects like the "Science Data Lake" have used embedding-based techniques to align OpenAlex topics with EDAM and MeSH. Direct perfect mapping is constrained by the different semantic axes each system uses.
4. bio.tools Registry
4.1 Overview and Statistics
bio.tools is the ELIXIR Tools & Data Services Registry, a comprehensive, community-driven catalog for bioinformatics software and data resources.
Statistics (May 2026):
| Metric | Value |
|--------|-------|
| Total entries | 33,010 |
| Total metadata annotations | 564,000+ |
| EDAM ontology annotations | 211,000+ |
| Contributors | 15,595 |
4.2 Registration Process
1. Create account: Users register at bio.tools
2. Add content: Fill structured metadata via web interface or API
3. Assign biotoolsID: Each tool receives a human-readable, URL-safe identifier (e.g.,
biotools:signalp)
4. EDAM annotation: Add topic, operation, data type, and format terms
5. Quality enrichment: Progressive curation following the Curators Guide
Registration methods:
Manual curation via web interface
API submission via REST endpoints
Literature mining via Pub2Tools (automated extraction from publications)
LLM-aided scoring followed by manual revision
4.3 REST API
Base URL: https://bio.tools/api/
Key Endpoints
| Method | Endpoint | Purpose |
|--------|----------|---------|
|
GET | /api/tool/ (or /api/t/) | List and search tools |
| POST | /api/tool/ | Register new tool (auth required) |
| PUT | /api/tool/ | Update existing tool (auth required) |
| DELETE | /api/tool/ | Delete tool (auth required) |
| POST | /api/rest-auth/login/ | Obtain auth token |
| GET | /api/rest-auth/user/ | Get authenticated user info |
Search Parameters
| Parameter | Purpose | Example |
|-----------|---------|---------|
|
q | General query (matches all attributes) | ?q=alignment |
| name | Search by tool name | ?name=signalp |
| topic | Filter by EDAM topic term | ?topic=Proteomics |
| topicID | Filter by EDAM topic URI/ID | ?topicID=topic_3510 |
| description | Search within descriptions | ?description=variant |
| version | Exact version match | ?version=5.0 |
| homepage | Exact homepage URL match | ?homepage=https://... |
| page | Result page number (default: 1) | ?page=2 |
| per_page | Results per page (default: 50) | ?per_page=100 |
| sort | Sort field | ?sort=lastUpdate |
| ord | Sort order | ?ord=desc |
Sort options:
lastUpdate, additionDate, name, score.
Example API Calls
`bash
Search for tools related to Proteomics
GET https://bio.tools/api/t/?topic=Proteomics&format=json
Search for tools performing sequence alignment
GET https://bio.tools/api/t/?q=sequence+alignment&format=json
Get details for a specific tool
GET https://bio.tools/api/tool/signalp/?format=json
`
4.4 Data Formats
The API supports:
JSON (default): For machine-readable access
XML: For legacy system compatibility
All data follows the biotoolsSchema, including EDAM annotations for operations, data types, formats, and topics.
5. Ecosystem Integration
5.1 ELIXIR Infrastructure
bio.tools, biotoolsSchema, and EDAM form a foundational pillar of the ELIXIR research infrastructure. The integration extends to:
Galaxy: Maps bio.tools EDAM annotations to internal tool metadata
CWL: ToolDog generates CWL descriptors from bio.tools entries
Nextflow: nf-core modules can reference bio.tools identifiers
WorkflowHub: Uses EDAM for workflow classification
5.2 Workflow System Integration
`mermaid
graph TB
subgraph "Standards Layer"
EDAM[EDAM Ontology]
BTS[biotoolsSchema]
end
subgraph "Registry Layer"
BT[bio.tools Registry]
WFH[WorkflowHub]
end
subgraph "Execution Layer"
GAL[Galaxy]
CWL[CWL Runners]
NF[Nextflow/nf-core]
end
EDAM --> BTS
BTS --> BT
EDAM --> WFH
BT --> GAL
BT --> CWL
BT --> NF
BT --> WFH
`
5.3 Automated Descriptor Generation
Tools like ToolDog leverage bio.tools metadata to automatically generate or enrich tool description templates for workflow languages:
1. Query bio.tools for tool metadata
2. Extract EDAM annotations (operations, inputs, outputs)
3. Generate CWL/Galaxy tool descriptors with pre-filled type information
4. Developers complete execution-specific details (command line, arguments)
6. Integration with Cytognosis
6.1 Mapping to Existing Schemas
The Cytognosis platform uses LinkML schemas in the
cytos knowledge graph. Several existing schemas directly overlap with biotoolsSchema/EDAM concepts:
| Cytognosis Schema | biotoolsSchema / EDAM Equivalent |
|------------------|----------------------------------|
|
cytos-topic (ResearchTopic) | EDAM Topic branch + OpenAlex + MeSH |
| cytos-core (DatasetRegistryEntry) | biotoolsSchema Summary + Labels groups |
| cytos-registries (plug-in registries) | bio.tools registry pattern |
| cytos-annotation (Annotation) | bio.tools curation model |
| ModalityEnum (phenotype, genotype, sc_omics, biosignal, imaging) | EDAM Data branch (partially) |
| ParserEnum (format handling) | EDAM Format branch |
6.2 EDAM as Skill-Tagging Ontology
The existing Cytognosis skill-tagging system uses a two-axis approach:
Axis A (coding/use-case): cyto:se/* — SWEBOK-derived
Axis B (org function): apqc:pcf:* — APQC PCF
EDAM could serve as a third axis for scientific operations:
Axis C (scientific function): edam:operation/* — EDAM operations
This maps directly to the cytoskills system. Each skill could declare:
Which EDAM operations it performs
Which EDAM data types it consumes/produces
Which EDAM formats it supports
Which EDAM topics it relates to
6.3 biotoolsSchema for the LEGO Model Registry
The Cytognosis "LEGO biological model" concept (composable computational biology modules) can adopt biotoolsSchema patterns:
1. Function model: Each LEGO block declares its operations, inputs, and outputs using EDAM terms
2. Composability: EDAM data type matching enables automatic compatibility checking between blocks
3. Discovery: EDAM topic classification enables researchers to find relevant blocks
4. Provenance: biotoolsSchema credit model tracks contributors
`yaml
Example: Cytognosis LEGO block described with biotoolsSchema concepts
lego_block:
name: "CytoVCF Variant Annotator"
description: "Annotates VCF variants with clinical significance"
function:
- operation:
- uri: edam:operation_3661 # SNP annotation
- uri: edam:operation_3672 # Gene functional annotation
input:
- data: edam:data_3498 # Sequence variations
format: edam:format_3016 # VCF
output:
- data: edam:data_3498 # Sequence variations (annotated)
format: edam:format_3016 # VCF
topic:
- edam:topic_3325 # Rare diseases
- edam:topic_3577 # Personalised medicine
toolType: library
language: Python
license: Apache-2.0
`
6.4 Experiment Orchestration
biotoolsSchema's function model maps to our experiment orchestration system:
| Experiment System Concept | biotoolsSchema Equivalent |
|--------------------------|---------------------------|
| Pipeline step | biotoolsSchema Function |
| Step operation | EDAM Operation |
| Step input data type | EDAM Data + Format |
| Step output data type | EDAM Data + Format |
| Pipeline topic area | EDAM Topic |
This enables:
Automatic type checking: Verify that one step's output format matches the next step's input format
Pipeline discovery: Find relevant pipeline steps by EDAM operation
Interoperability: Exchange pipeline definitions with Galaxy, CWL, Nextflow ecosystems
6.5 Should We Register Tools in bio.tools?
Recommendation: Yes, after MVP launch.
Benefits:
Visibility in the ELIXIR ecosystem
Discoverability by 15,000+ bio.tools contributors
Standardized citation via biotoolsID
Integration with Galaxy, CWL, WorkflowHub
FAIR software compliance
Timing: Register after our first public tools are stable. Start with:
1. Cytos KG query tools
2. LEGO biological model blocks
3. Experiment orchestration pipeline components
6.6 Relationship to Topic Schema
The existing
cytos-topic schema integrates OpenAlex (4-level hierarchy), OpenAlex Concepts (6-level tree), UMLS Semantic Network, MeSH, and ASJC. EDAM Topics represent a complementary axis:
`
Existing topic sources (cytos-topic):
├── openalex_topic (4,516 topics)
├── openalex_concept (65K+ concepts)
├── mesh (NLM Subject Headings)
├── umls_semantic_network (UMLS types)
├── asjc (Elsevier ASJC)
└── custom (Cytos-defined)
Proposed addition:
└── edam_topic (~3,472 terms, bioinformatics-specific)
`
Adding EDAM as a TopicSource would provide precise bioinformatics-specific classification complementing the broader MeSH and OpenAlex coverage.
7. Recommendations
Immediate Actions (Phase 0)
1. Add EDAM to TopicSource enum in
cytos-topic.yaml:
`yaml
edam_topic:
description: "EDAM ontology bioinformatics topic."
edam_operation:
description: "EDAM ontology operation type."
`
2. Generate EDAM validation enum analogous to existing
efo_validation_enum.yaml and mondo_validation_enum.yaml` for EDAM terms
3. Design LEGO block metadata schema using biotoolsSchema Function model as template, implemented in LinkML
4. Implement EDAM-based skill tagging as Axis C in the cytoskills system
5. Build bio.tools API client in cytoskeleton for querying existing tools
6. Create mapping file between EDAM Topics and existing OpenAlex/MeSH terms in cytos-topic
7. Register Cytognosis tools in bio.tools after public release
8. Integrate EDAM type checking into experiment orchestration pipeline validation
9. Generate CWL descriptors from LEGO block biotoolsSchema metadata via ToolDog patterns