Docs/BioAgents Architecture
Documentation

BioAgents Architecture

BioAgents is an AI powered system that processes research data, extracts metadata, identifies entities, and generates knowledge graphs to enhance data discoverability and usability.

System Overview

BioAgents provides intelligent processing capabilities for research data within the Bio-DID-Seq ecosystem. It uses large language models (LLMs) and specialized AI agents to extract valuable information from research papers, identify entities, generate knowledge graphs, and enable natural language querying of research data.

INTERFACEResearcher InterfaceBio-DID-Seq Core PlatformAGENT FRAMEWORKAgent RuntimePlugin SystemTask WorkersMemory ManagementKNOWLEDGE PROCESSINGDocument ProcessingEntity RecognitionRelationship DetectionGraph GenerationKNOWLEDGE STORAGERDF Triple StoreJSON-LD DocumentsEntity DatabaseSearch IndexINTEGRATION LAYERDID IntegrationUCAN AuthorizationIPFS Storagedispatch tasksextract knowledgepersistconnect
BioAgents Architecture

Core Components

1. Agent Framework

BioAgents is built on the Eliza OS agent framework which provides:

  • Agent Runtime: Core execution environment for AI agents
  • Plugin System: Extensible architecture for specialized capabilities
  • Memory Management: Persistent storage of agent knowledge and states
  • Task Workers: Async processing of compute intensive tasks

2. Knowledge Processing Pipeline

The BioAgents system processes research data through a multi stage pipeline:

Document Processing

  • PDF Extraction: Extracts text and structure from PDFs using GROBID
  • Content Normalization: Converts different formats to structured text
  • Section Identification: Recognizes abstract, methods, results, etc.

Knowledge Extraction

  • Entity Recognition: Identifies genes, proteins, diseases, compounds, etc.
  • Relationship Detection: Finds connections between biological entities
  • Claim Extraction: Identifies key scientific claims and evidence
  • Metadata Collection: Authors, citations, publication details, etc.

Knowledge Representation

  • RDF Generation: Converts extracted information to Resource Description Framework
  • JSON-LD Creation: Structured, interoperable representation of knowledge
  • Knowledge Graph Integration: Links new knowledge with existing data

Integration with Bio-DID-Seq

1. API Integration

Upload Research PaperProcess Paper RequestExtract KnowledgeGenerate Knowledge GraphStore KnowledgeReturn ResultsDisplay ResultsResearcherBio-DID-Seq APIBioAgents ServiceKnowledge Graph
API Integration

2. DID based metadata management

Request with UCAN TokenVerify UCAN CapabilitiesForward RequestQuery Knowledge with DIDReturn Authorized DataReturn ResultsDisplay ResultsUserBio-DID-SeqBioAgentsKnowledge Graph
DID based metadata management

3. Data Processing Flow

INGESTResearch PaperuploadBio-DID-Seq APIIPFS StoragestoreAI PROCESSINGBioAgentsextract · identify · generateMetadataBiological EntitiesKnowledge GraphSTORAGE & INDEXBio-DID RegistryDID metadataEntity DatabaseRDF Triple StoreSearch IndexDISCOVERYSPARQL EndpointDiscovery Interfaceprocessindexquery
Data Processing Flow

API Specifications

BioAgents API Endpoints

EndpointMethodDescription
/api/bioagents/processPOSTProcess a research paper
/api/bioagents/statusPOSTCheck processing status
/api/bioagents/metadataPOSTGet extracted metadata
/api/bioagents/searchPOSTSearch biological entities
/api/bioagents/knowledge-graphPOSTGenerate knowledge graph
/api/bioagents/queryPOSTQuery using natural language

Request/Response Examples

Process Paper Request

json
{
  "file_cid": "QmXg9Pp2ytZ14xgK35M6iTC2Vz6jR9zYgooNp2UHPTMnPN",
  "title": "CRISPR-Cas9 Gene Editing for Neurodegenerative Diseases",
  "authors": ["Jane Smith", "John Doe"],
  "doi": "10.1038/s41586-021-03819-2"
}

Process Paper Response

json
{
  "task_id": "b8e5c9a4-2c7a-4d64-b4b3-7c01e2f8b54e",
  "status": "processing"
}

Security and Privacy

BioAgents maintains the same security and privacy standards as the Bio-DID-Seq system:

  • Authentication: All API calls require valid authentication
  • Authorization: UCAN based capability model for granular permissions
  • Data Encryption: All data in transit and at rest is encrypted
  • Audit Logging: All data access is logged for compliance
  • Privacy Controls: Data processors follow GDPR principles
Bio-DID-SequencerKnowledge Graph