🔬 Eldric Science Worker

AI-powered scientific computation engine combining bioinformatics, CRISPR design, molecular docking, and laboratory management with multi-backend LLM integration for intelligent research assistance.

eldric-scienced v4.0.0 · Port 8897 🤖 AI-Assisted Research

🚀 New: AI-Assisted Research

The Science Worker now integrates with any LLM backend (Ollama, vLLM, OpenAI, Anthropic) for AI-assisted analysis. Generate hypotheses, interpret results, design experiments, and delegate complex research tasks to the Agent Worker's multi-agent system.

What Can You Do?

🧬

Analyze Sequences

Upload DNA, RNA, or protein sequences for comprehensive analysis including GC content, molecular weight, translation, and BLAST search.

✂️

Design CRISPR Guides

Generate optimized guide RNAs for Cas9/Cas12a with off-target prediction, efficiency scoring, and base/prime editing support.

💊

Discover Drugs

Perform molecular docking, predict ADMET properties, and analyze drug-target interactions with GPU acceleration.

🏢

Manage Lab Data

Track samples, experiments, and protocols with full regulatory compliance (FDA 21 CFR Part 11, GLP, GMP).

🤖

AI Analysis

Use LLMs to interpret results, generate hypotheses, suggest experiments, and write research summaries automatically.

🔗

Access 62 APIs

Query NASA, CERN, NCBI, UniProt, and 62 scientific databases through the integrated Data Worker connector.

Architecture

Science Worker - AI-Integrated Research Platform

🔬 Research Agent

The Science Worker operates as an intelligent Research Agent that orchestrates federated searches across multiple scientific data providers. Rather than making direct API calls, it delegates to the Data Worker's 62 scientific connectors and synthesizes results into comprehensive research reports with citations.

🌐 Federated Search Architecture

When you conduct research through the Science Worker, it automatically queries multiple providers simultaneously (arXiv, Crossref, OpenAlex, PubMed, Semantic Scholar, etc.), normalizes the heterogeneous responses, extracts citations, and synthesizes findings into a unified research report. The Data Worker handles rate limiting, caching, and API authentication for all 62 providers.

Federated Research Pipeline

Research Agent Endpoints

POST /api/v1/research/search

Federated search across multiple scientific providers

POST /api/v1/research/conduct

Full research workflow: search → synthesize → report

POST /api/v1/research/plan

Generate a research plan with suggested providers

GET /api/v1/research/providers

List available providers grouped by topic

Example: Conduct Research

curl -X POST http://localhost:8897/api/v1/research/conduct \
  -H "Content-Type: application/json" \
  -d '{
    "query": "machine learning",
    "topic": "ai",
    "providers": ["arxiv", "crossref", "openalex", "pubmed", "semanticscholar"],
    "max_results_per_provider": 5
  }'

# Response:
{
  "query": "machine learning",
  "summary": "Found 10 sources across 4 providers. AI synthesis unavailable...",
  "citations": [
    {
      "title": "Scikit-learn: Machine Learning in Python",
      "authors": "Pedregosa et al.",
      "date": "2011",
      "id": "10.5555/1953048.2078195",
      "provider": "crossref",
      "snippet": "Scikit-learn is a Python module integrating..."
    },
    ...
  ],
  "key_findings": ["Machine learning with sklearn (crossref)", ...],
  "sources_consulted": 5,
  "sources_cited": 10,
  "generated_at": "2026-02-01T15:13:56Z"
}

Supported Research Topics

📚 Literature

arXiv, Crossref, OpenAlex, PubMed, Semantic Scholar, Europe PMC

🧬 Life Sciences

NCBI, UniProt, Ensembl, PDB, ChEMBL, MyGene, ClinicalTrials

🌍 Earth Sciences

NOAA, USGS, Copernicus, GBIF, OBIS, iNaturalist, eBird

⚛️ Physics

CERN, GWOSC, INSPIRE-HEP, NASA, ESA, MAST

🔬 Chemistry

PubChem, COD, Materials Project, NIST

💊 Pharma

OpenFDA, DrugBank, ChEMBL, ClinicalTrials

🤖 AI-Assisted Research

The Science Worker integrates with multiple LLM backends to provide intelligent research assistance. Use AI for basic tasks directly, or delegate complex multi-step research to the Agent Worker.

AI Research Workflow

💬

Direct LLM Access

For quick AI tasks, query LLMs directly through any configured inference backend.

Text completion for research summaries
Multi-turn chat for interactive analysis
Data interpretation and summarization
Hypothesis generation with confidence scores
CRISPR guide evaluation with AI scoring
Drug target prediction assistance
Experiment suggestion with protocols

Supported Backends

Ollama

vLLM

OpenAI

Anthropic

TGI

llama.cpp

🔗

Agent Worker Delegation

For complex multi-step tasks, delegate to the Agent Worker's multi-agent system.

Complex research task delegation
Automated literature reviews (up to 20+ papers)
Multi-agent collaboration (Researcher + Validator)
Protocol generation with constraints
Agentic RAG with iterative retrieval
Query decomposition for complex questions
Knowledge base integration

Agent Types

Researcher

Validator

Planner

Coder

AI API Endpoints

GET /api/v1/ai/status

Check LLM and Agent Worker availability

POST /api/v1/ai/completion

Generate text completion for research tasks

POST /api/v1/ai/chat

Multi-turn chat for interactive research

POST /api/v1/ai/analyze

AI analysis: summarize, interpret, suggest, compare

POST /api/v1/ai/hypothesis

Generate research hypotheses with confidence scores

POST /api/v1/ai/experiments

Get experiment suggestions with protocols

POST /api/v1/agent/research

Delegate complex research to Agent Worker

POST /api/v1/agent/literature

Request automated literature review

POST /api/v1/agent/multi-analysis

Multi-agent collaborative analysis

POST /api/v1/agent/protocol

Generate experimental protocols

Computational Features

🧬

Bioinformatics Engine

DNA/RNA/protein sequence analysis with built-in algorithms for molecular biology research.

Sequence analysis (GC content, molecular weight, melting temp)
DNA translation (6-frame, codon optimization)
BLAST alignment (local database or NCBI)
Multiple sequence alignment (Clustal, MUSCLE)
Variant calling and annotation (VCF)
Phylogenetic tree construction

Integrated Tools

BLAST+, HMMER, Clustal Omega, MUSCLE, SAMtools, BCFtools

✂️

CRISPR Design

Guide RNA design with off-target analysis and efficiency prediction for genome editing.

Cas9/Cas12a guide RNA design
Off-target site prediction (CFD/MIT scoring)
On-target efficiency prediction (Rule Set 2)
Base editing guide design (ABE/CBE)
Prime editing (pegRNA design)
HDR template generation

Algorithms

Doench scoring, CFD scoring, Lindel prediction, DeepCRISPR

💊

Drug Discovery

Molecular docking, property prediction, and structure analysis for pharmaceutical research.

Molecular docking (AutoDock Vina, GNINA)
ADMET property prediction
Lipinski Rule of Five / Veber rules
SMILES/InChI parsing and validation
Molecular fingerprints (ECFP, MACCS)
AlphaFold structure prediction integration

Integrated Tools GPU

AutoDock Vina, GNINA, RDKit, Open Babel, ESMFold

🏢

LIMS (Laboratory Management)

Laboratory Information Management with regulatory compliance for research environments.

Sample registration and barcode tracking
Experiment workflow management
Equipment calibration tracking
Reagent inventory management
Audit trails (21 CFR Part 11 compliant)
GLP/GMP compliance modes

Compliance

FDA 21 CFR Part 11, EU Annex 11, ISO 17025, GLP, GMP

REST API

Bioinformatics

POST /api/v1/bio/analyze

Analyze DNA/RNA/protein sequence (GC%, MW, Tm)

POST /api/v1/bio/translate

Translate DNA to protein (all reading frames)

POST /api/v1/bio/blast

BLAST search against local or NCBI databases

POST /api/v1/bio/align

Multiple sequence alignment

POST /api/v1/bio/variants

Variant calling from alignment

CRISPR

POST /api/v1/crispr/design

Design guide RNAs for target sequence

POST /api/v1/crispr/offtargets

Predict off-target sites for guide

POST /api/v1/crispr/base-edit

Design base editing guides (ABE/CBE)

POST /api/v1/crispr/prime-edit

Design prime editing guides (pegRNA)

Drug Discovery

POST /api/v1/pharma/compound

Parse SMILES, calculate molecular properties

POST /api/v1/pharma/dock

Molecular docking (Vina/GNINA)

POST /api/v1/pharma/admet

ADMET property prediction

POST /api/v1/pharma/structure

Fetch/predict protein structure

LIMS

POST /api/v1/lims/samples

GET /api/v1/lims/samples/{id}

Get sample by ID or barcode

POST /api/v1/lims/experiments

Create experiment record

GET /api/v1/lims/audit/{type}/{id}

Get audit trail (compliance)

Quick Start

# Install Science Worker
sudo dnf install eldric-scienced-4.0.0-1.el9.x86_64.rpm

# Start daemon with AI support
sudo systemctl start eldric-scienced

# Or run manually with all integrations
eldric-scienced --port 8897 \
  --controller http://controller:8880 \
  --inference-workers http://localhost:8890 \
  --agent-workers http://localhost:8893 \
  --data-workers http://localhost:8892 \
  --gpu --cuda-devices 0,1

Example: AI Analysis of BLAST Results

# First, run a BLAST search
curl -X POST http://localhost:8897/api/v1/bio/blast \
  -H "Content-Type: application/json" \
  -d '{"sequence": "ATGCGATCG...", "database": "nr"}'

# Then analyze results with AI
curl -X POST http://localhost:8897/api/v1/ai/analyze \
  -H "Content-Type: application/json" \
  -d '{
    "data": { "blast_results": [...] },
    "type": "interpret"
  }'

# Response:
{
  "analysis_type": "interpret",
  "analysis": "The BLAST results show significant homology to...",
  "model": "llama3.1:70b"
}

Example: Generate Research Hypothesis

curl -X POST http://localhost:8897/api/v1/ai/hypothesis \
  -H "Content-Type: application/json" \
  -d '{
    "context": "We observed increased expression of BRCA1 in tumor samples
                compared to normal tissue. The expression correlated with
                patient age but not tumor stage..."
  }'

# Response includes ranked hypotheses with confidence scores

Example: Automated Literature Review

curl -X POST http://localhost:8897/api/v1/agent/literature \
  -H "Content-Type: application/json" \
  -d '{
    "topic": "CRISPR-Cas9 off-target effects in therapeutic applications",
    "max_papers": 20
  }'

# Agent Worker searches PubMed, arXiv, bioRxiv and generates a comprehensive review

Configuration

{
  "port": 8897,
  "controller_url": "http://controller:8880",
  "data_worker_urls": ["http://localhost:8892"],
  "inference_worker_urls": ["http://localhost:8890"],
  "agent_worker_urls": ["http://localhost:8893"],

  "llm": {
    "default_model": "llama3.2:3b",
    "research_model": "llama3.1:70b",
    "max_tokens": 4096,
    "temperature": 0.7,
    "enable_streaming": false,
    "timeout_ms": 120000
  },

  "bioinformatics_backend": "default",
  "blast_databases_path": "/data/blast_db",
  "pharma_backend": "default",
  "pdb_cache_path": "/data/pdb_cache",
  "lims_database_path": "/data/lims.db",
  "compliance_mode": "fda_21_cfr_11"
}

License Tiers

Feature	Free	Standard	Professional	Enterprise
Sequence Analysis	✓	✓	✓	✓
AI Analysis	Basic	✓	✓	✓
Agent Delegation	✗	✓	✓	✓
CRISPR Design	5/day	50/day	500/day	Unlimited
BLAST Search	✗	✓	✓	✓
Molecular Docking	✗	✗	✓	✓
AlphaFold Integration	✗	✗	✗	✓
LIMS Samples	10	100	1,000	Unlimited
21 CFR Part 11	✗	✗	✗	✓
GPU Acceleration	✗	✓	✓	✓
Science Workers	1	2	5	Unlimited

Port Reference

Component	Port	Description
Science Worker	`8897`	Scientific computation + AI research
Controller	`8880`	Cluster management, licensing
Data Worker	`8892`	Databases, scientific APIs (62), RAG storage
Inference Worker	`8890`	LLM inference (Ollama, vLLM, OpenAI...)
Agent Worker	`8893`	Multi-agent tasks, agentic RAG

🚀 New: AI-Assisted Research

What Can You Do?

Analyze Sequences

Design CRISPR Guides

Discover Drugs

Manage Lab Data

AI Analysis

Access 62 APIs

Architecture

🔬 Research Agent

🌐 Federated Search Architecture

Research Agent Endpoints

Example: Conduct Research

Supported Research Topics

📚 Literature

🧬 Life Sciences

🌍 Earth Sciences

⚛️ Physics

🔬 Chemistry

💊 Pharma

🤖 AI-Assisted Research

Direct LLM Access

Supported Backends

Agent Worker Delegation

Agent Types

AI API Endpoints

Computational Features

Bioinformatics Engine

Integrated Tools

CRISPR Design

Algorithms

Drug Discovery

Integrated Tools GPU

LIMS (Laboratory Management)

Compliance

REST API

Bioinformatics

CRISPR

Drug Discovery

LIMS

Quick Start

Example: AI Analysis of BLAST Results

Example: Generate Research Hypothesis

Example: Automated Literature Review

Configuration

License Tiers

Port Reference

Related Documentation

Data Worker

Agent Worker

LLM Backends