genos-client 0.1.1

Genos API client for accessing genomic analysis services

Genos

Get API key | Quick start

Genos is a Python SDK for accessing genomic analysis models through the GeneOS platform. It provides a unified interface for variant pathogenicity prediction, DNA sequence embedding extraction, RNA-seq coverage track prediction, and genomic visualization.

Features

• 🧬 Variant Prediction: Assess pathogenicity of genetic variants
• 🔬 Embedding Extraction: Extract deep learning embeddings from DNA sequences
• 🧪 RNA-seq Coverage Track Prediction: Predict RNA-seq coverage tracks from genomic coordinates
• 📊 Genomic Visualization: Plot and analyze genomic tracks
• 🚀 Easy to Use: Simple, intuitive API with comprehensive error handling
• 🔐 Robust Authentication: Automatic token validation and payment checking
• ⚠️ Comprehensive Error Handling: Specific exceptions for different error types

Get API Key

To use Genos, you need an API key from the DCS Cloud.

Request Access

1. Log in to the DCS Clould and navigate to Personal Center → API Key Management.
2. Click “Create API Key”.
3. Read the “API Usage Notice” and confirm your agreement.
4. The system will automatically generate your exclusive API Key — copy and keep it safe.

⚠️ Please store your key securely and avoid any unauthorized disclosure.

Usage Policy

• The API Key is for personal use only — sharing, transferring, or publishing is strictly prohibited.
• If the key is leaked, misused, or used for illegal purposes, the platform reserves the right to immediately disable it.
• You can manually deactivate your key in the control panel at any time.

Installation

Install from Source

git clone https://github.com/BGI-HangzhouAI/Genos.git
cd sdk
pip install -e .

Install from Pypi

pip install genos-client

Requirements

• Python 3.10 or higher
• pip package manager

Quick Start

Basic Usage

from genos import create_client

# Create client (uses GENOS_API_TOKEN environment variable)
client = create_client()

# Or provide token explicitly
client = create_client(token="your_api_token_here")

1. Variant Pathogenicity Prediction

Predict whether a genetic variant is pathogenic or benign:

# Predict variant pathogenicity
result = client.variant_predict("hg19", "chr6", 51484075, "T", "G")['result']

print(f"Variant: {result['variant']}")
print(f"Prediction: {result['prediction']}")
print(f"Pathogenic Score: {result['score_Pathogenic']:.4f}")
print(f"Benign Score: {result['score_Benign']:.4f}")

2. DNA Sequence Embedding Extraction

Extract deep learning embeddings from DNA sequences:

# Extract embedding for a single sequence
sequence = "ATCGATCGATCGATCGATCGATCGATCG"
result = client.get_embedding(sequence, model_name="Genos-1.2B")['result']

print(f"Sequence Length: {result['sequence_length']}")
print(f"Embedding Dimension: {result['embedding_dim']}")
print(f"Embedding Shape: {result['embedding_shape']}")

# Access the embedding vector
embedding_vector = result['embedding']  # List of floats

Available Models:

• Genos-1.2B: 1.2 billion parameter model
• Genos-10B: 10 billion parameter model

Pooling Methods:

• mean: Average pooling across sequence
• max: Max pooling
• last: Use last token embedding
• none: Return all token embeddings

3. RNA-seq Coverage Track Prediction

Predict RNA-seq coverage tracks based on genomic coordinates:

# Predict RNA-seq coverage track
result = client.rna_coverage_track_pred(chrom="chr6", start_pos=51484075)['result']

print(f"Predicted coverage track: {result}")

Advanced Configuration

Custom Embedding Service

GenosClient allows users to configure a custom embedding API endpoint.
This is useful if you want to deploy your own embedding service locally or within your organization.

Note: Now, the variant prediction and RNA-seq coverage track prediction models are not open-source and cannot be self-hosted. Only the embedding service can be customized.

from genos import GenosClient

# Initialize the client with a custom embedding endpoint
client = GenosClient(
    token="your_custom_token",  # Your token for authenticating with your own embedding service
    api_map={
        # Only the embedding service can be customized
        "embedding": "https://custom-embed-api.example.com/predict"
    }
)

# Calls to variant and RNA APIs will still use the official hosted services

Timeout Configuration

Adjust request timeout for long-running operations:

# Set 60-second timeout
client = create_client(token="your_token", timeout=60)

Error Handling

Genos provides comprehensive error handling with specific exception types for different scenarios. All API responses follow a consistent format.

Error Response Format

All error responses from the Genos API follow this structure:

{
  "result": {},
  "status": "<HTTP_STATUS_CODE>",
  "messages": "<ERROR_MESSAGE>"
}

Common Error Codes

Status Code	Error Message	Description
400	Insufficient balance	Your account balance is insufficient for the requested operation
401	Invalid API Key	The provided API key is invalid or expired
500	Internal server error	An unexpected error occurred on the server side

Examples

Complete examples are available in the examples/ directory:

• predict_variant.py: Variant pathogenicity prediction
• embedding_extract.py: DNA Sequence Embedding Extraction
• rna_generator.py: RNA-seq Coverage Track Prediction
• error_handling_demo.py: Comprehensive error handling examples

Primary Paper

Genos: A Human-Centric Genomic Foundation Model

Adi Lin¹, Bin Xie¹, Cheng Ye¹,Cheng Wang¹, Duoyuan Chen¹, Ercheng Wang¹, Fanfeng Lu¹, Guirong Xue¹, Haiqiang Zhang¹, Jiajie Zhan¹, Jianfeng Zhang¹, Jiangshuan Pang¹, Jianqiang Liang¹, Jiawei Lin¹, Jiaxin Ma¹, Jie Hu¹, Jing Ma¹,Jingni Dong¹, Jiongzhen Li, Junchen Liu¹, Junhong Chen¹, Junyou Li¹, Kai Ding¹, Kaiwen Deng¹, Kui Chen¹, Lihui Wang¹, Longqi Liu¹, Ling Guo¹, Liwen Xiong¹, Luhao Yang¹, Ming Cheng¹, Nanning Chen¹, Renzhong Chen¹, Shanxin Sun¹, Shaoshuai Li¹, Shicheng Chen¹, Shiping Liu¹, Siwei Xie¹, Suyan Liu¹, Tao Zhou¹, Wangyang Tang¹, Weiqiang Zhang¹, Xianyue Jiang¹, Xianzhi Qi¹, Xin Jin¹, Xinjiang Tan¹, Xinyue Hu¹, Xun Xu¹, Xuyang Feng¹, Yafei Lu¹ ,Yifan Gao¹, Yong Shang¹, Youzhe He¹, Yue Yuan¹, Yufan Wang¹, Yuqi Liu1, Zhan Xiao¹, Zhangyuan Meng¹, Zhaorong Li¹, Zhe Zhao¹, Zheng Yang¹, Zilin Wang¹

GigaScience, 2025.
https://doi.org/10.1093/gigascience/giaf132

How to Cite

If you use the SDK, please include the following citation in your work:

@article{he2025genos,
  title   = {Genos: A Human-Centric Genomic Foundation Model},
  author  = {Adi Lin¹, Bin Xie¹, Cheng Ye¹,Cheng Wang¹, Duoyuan Chen¹, Ercheng Wang¹, Fanfeng Lu¹, Guirong Xue¹, Haiqiang Zhang¹, Jiajie Zhan¹, Jianfeng Zhang¹, Jiangshuan Pang¹, Jianqiang Liang¹, Jiawei Lin¹, Jiaxin Ma¹, Jie Hu¹, Jing Ma¹,Jingni Dong¹, Jiongzhen Li, Junchen Liu¹, Junhong Chen¹, Junyou Li¹, Kai Ding¹, Kaiwen Deng¹, Kui Chen¹, Lihui Wang¹, Longqi Liu¹, Ling Guo¹, Liwen Xiong¹, Luhao Yang¹, Ming Cheng¹, Nanning Chen¹, Renzhong Chen¹, Shanxin Sun¹, Shaoshuai Li¹, Shicheng Chen¹, Shiping Liu¹, Siwei Xie¹, Suyan Liu¹, Tao Zhou¹, Wangyang Tang¹, Weiqiang Zhang¹, Xianyue Jiang¹, Xianzhi Qi¹, Xin Jin¹, Xinjiang Tan¹, Xinyue Hu¹, Xun Xu¹, Xuyang Feng¹, Yafei Lu¹ ,Yifan Gao¹, Yong Shang¹, Youzhe He¹, Yue Yuan¹, Yufan Wang¹, Yuqi Liu1, Zhan Xiao¹, Zhangyuan Meng¹, Zhaorong Li¹, Zhe Zhao¹, Zheng Yang¹, Zilin Wang¹},
  journal = {GigaSience},
  year    = {2025},
  doi     = {10.1093/gigascience/giaf132}
}