🧬 OpenGenomeLLM-70B just released — State-of-the-art on GeneTuring benchmark  ·  Download now →

📚 Documentation

Everything you need to get started with OpenGenomeLLM — quickstart, API reference, and model guides.

🚀 Quickstart

Welcome to OpenGenomeLLM — the open-source platform for genomic AI. This guide will get you up and running in under 5 minutes.

New: OpenGenomeLLM-70B achieves 78.9% on GeneTuring benchmark, surpassing GPT-4 on genomic tasks. See models →
⚡
Installation
Install the Python SDK and get set up in minutes.
🔌
REST API
Use the HTTP API from any language or tool.
🧬
VCF Interpretation
Analyze VCF files and generate clinical reports.
🎯
Fine-tuning
Adapt models to your institution's data and needs.

Installation

Install the OpenGenomeLLM Python SDK using pip:

bash
pip install opengenomellm

# Or install from source for the latest features
git clone https://github.com/deepcog-ai/opengenomellm
cd opengenomellm && pip install -e .

Python SDK — Basic Usage

Load and run the OpenGenomeLLM-70B model for variant interpretation:

python
from opengenomellm import GenomicLLM

# Initialize the model
model = GenomicLLM("deepcog-ai/OpenGenomeLLM-70B")

# Interpret a variant
result = model.interpret_variant(
    gene="BRCA1",
    variant="c.5266dupC",
    population="South Asian"
)

print(result.pathogenicity)   # "Pathogenic"
print(result.acmg_class)      # "Class V"
print(result.clinical_report) # Full narrative report

VCF File Interpretation

Process an entire VCF file and generate a comprehensive clinical report:

python
from opengenomellm import GenomicLLM, VCFProcessor

model = GenomicLLM("deepcog-ai/OpenGenomeLLM-70B")
processor = VCFProcessor()

# Load and process VCF
variants = processor.load("patient_001.vcf")
report = model.generate_clinical_report(
    variants=variants,
    patient_info={"age": 45, "sex": "F", "ethnicity": "South Asian"},
    format="pdf"
)

report.save("clinical_report_001.pdf")

REST API

Use the OpenGenomeLLM REST API from any programming language. All endpoints require an API key from your account dashboard.

bash
curl -X POST https://api.opengenomellm.org/v1/interpret \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "opengenomellm-70b",
    "gene": "BRCA1",
    "variant": "c.5266dupC",
    "population": "South Asian"
  }'

Docker Deployment

Deploy OpenGenomeLLM-70B in your own infrastructure using our official Docker image:

bash
# Pull and run with GPU support
docker pull deepcogai/opengenomellm:70b-latest

docker run --gpus all -p 8000:8000 \
  -e MODEL=opengenomellm-70b \
  -e QUANTIZATION=4bit \
  deepcogai/opengenomellm:70b-latest

# API will be available at http://localhost:8000