Published · Nature Communications 2024

下一代AI分析
多组学生态

A unified framework for exploring bulk, single-cell, and spatial RNA-seq. From raw counts to publication-ready figures — in a single pipeline.

Launch Web App Documentation GitHub
pip install omicverse
v1.7.9
70+ Integrated Tools
3 Omics Domains
Nature Communications 2024
GPU Accelerated
1000+ Registered Functions
Capabilities

Everything You Need for Transcriptomics

A comprehensive analysis toolkit covering the full transcriptomics workflow — from raw data to publication-ready figures.

Bulk
Bulk RNA-seq Analysis
Complete pipeline: DESeq2-based differential expression, WGCNA co-expression networks, GSEA pathway enrichment, and PPI network construction.
DESeq2WGCNAGSEAPPITCGA
Single-cell
Single-Cell Analysis
End-to-end scRNA-seq: QC, normalization, clustering, trajectory inference, cell-cell communication, and multi-modal integration.
SCSAMetaTiMEVIACellPhoneDBMOFA
Spatial
Spatial Transcriptomics
Supports Visium, Slide-seq, MERFISH, and seqFISH. Spatial clustering, deconvolution, trajectory analysis, and multi-slide alignment.
STAGATETangramSTTSpaceFlow
Multi-omics
Multi-Omics Integration
Bridge bulk and single-cell data with BulkTrajBlend (Beta-VAE + GNN). Spatial deconvolution and cross-modal paired integration.
BulkTrajBlendGLUEBulk2Single
AI
AI-Powered Analysis
Smart Agent with LLM integration drives analyses in natural language. Compatible with OpenAI, Claude, Gemini, and local models.
GPT-4oClaudeNatural LanguageFunction Registry
Viz
High-Performance Visualization
WebGL-accelerated rendering for 200K+ cells via deck.gl. Interactive Plotly charts and publication-ready matplotlib figures.
WebGLdeck.glPlotlyGPU
AI · bioRxiv 2026
OmicClaw
Executable and reproducible natural-language multi-omics analysis. Describe your analysis in plain language — OmicClaw generates, executes, and documents reproducible workflows automatically.
Natural LanguageReproducibleMulti-omics 📄 Paper
AI · Web
J.A.R.V.I.S
The built-in AI agent inside OmicVerse Web. Understands your biological context, selects the right tools, and guides you through the entire analysis interactively — from upload to final figure.
GPT-4oClaudeGeminiLocal LLMInteractive
API Modules

Organized by Workflow

Each module is purpose-built for a specific omics domain, with consistent APIs and over 1000 discoverable functions.

Module
ov.bulk
Bulk RNA-seq: DESeq2 DEG, WGCNA, GSEA, batch correction, PPI networks, TCGA integration
Module
ov.single
scRNA-seq: cell annotation, trajectory inference, CCC, drug response, regulatory networks
Module
ov.space
Spatial: clustering, deconvolution, multi-slide alignment, trajectory, spatially variable genes
Module
ov.pp
Preprocessing: QC, normalization, HVG selection, PCA/UMAP, leiden/louvain clustering
Module
ov.pl
Visualization: embedding, volcano, violin, dot, heatmap, spatial plots — publication ready
Module
ov.bulk2single
Multi-modal bridging: BulkTrajBlend, Beta-VAE generation, GNN interpolation, spatial mapping
Module
ov.agent
LLM Smart Agent: natural language analysis, function recommendation, auto-code generation
Module
ov.utils
Utilities: data I/O, color palettes, function registry — find_function, list_functions, recommend_function
Module
ov.external
Integrations: consistent wrappers for 70+ popular bioinformatics tools
Discoverable Function Registry

Find any function using natural language. Semantic search across 1000+ registered functions with examples and related recommendations.

# Discover functions by description or keyword
ov.find_function("cell type annotation")
ov.recommend_function("differential expression")
ov.list_functions(category="single")
Web Interface

Analyze Without Writing Code

OmicVerse Web brings the full library to an interactive browser interface — no Python knowledge required.

1
Upload H5AD / AnnData
Load your dataset directly from the file system. Preview mode for large files avoids full memory loading.
2
Interactive Analysis
Run clustering, DEG, trajectory, and cell annotation analyses through guided parameter panels on the left sidebar.
3
GPU-Accelerated Visualization
Explore UMAP / t-SNE embeddings with WebGL deck.gl — instant rendering for 200K+ cell datasets.
4
Export & Save
Save processed AnnData, download publication-ready figures, or export analysis as Jupyter notebooks.
Open Web App
Feature
Embedded Terminal
Full PTY terminal (xterm.js) for direct Python / shell access alongside the analysis UI
Feature
Notebook Editor
Create and execute Jupyter-style cells with CodeMirror Python syntax highlighting
Feature
AI Agent Chat
Built-in LLM agent for natural language analysis guidance and code generation
Feature
File Manager
Browse, upload, rename, and manage data files directly from the browser
Feature
Package Manager
Install conda/pip packages with mirror selection and real-time install output
Feature
Bilingual UI
Complete English and Chinese interface with one-click language toggle and dark mode
Get Started

Install OmicVerse

Available via pip, conda/mamba, uv, and Docker. Supports Python 3.8–3.12 on Linux, macOS (incl. Apple Silicon), and Windows.

Python 3.8, 3.9, 3.10, 3.11, 3.12
Linux · macOS · Windows (WSL)
Apple Silicon native support
Optional CUDA GPU acceleration
Conda-forge, PyPI, and Docker images
# Standard install pip install omicverse # With optional GPU support (PyTorch required) pip install omicverse[gpu]
# Create a dedicated environment conda create -n omicverse python=3.10 conda activate omicverse # Install with mamba (faster solver) mamba install -c conda-forge omicverse
# Ultra-fast install with uv uv pip install omicverse # In a virtual environment uv venv --python 3.11 uv pip install omicverse
# Install the development version git clone https://github.com/Starlitnightly/omicverse cd omicverse pip install -e .
Publication

Cite OmicVerse

If OmicVerse contributed to your research, please cite our Nature Communications paper.

Nature Communications · 15:5983 · 2024
OmicVerse: a framework for bridging and deepening insights across bulk and single-cell sequencing
Zehua Zeng, Yuqing Ma, Lu Hu, et al.
DOI: 10.1038/s41467-024-50194-3
View Paper
Show BibTeX 
@article{zeng2024omicverse,
  title     = {OmicVerse: a framework for bridging and deepening
               insights across bulk and single-cell sequencing},
  author    = {Zeng, Zehua and Ma, Yuqing and Hu, Lu and others},
  journal   = {Nature Communications},
  volume    = {15},
  pages     = {5983},
  year      = {2024},
  doi       = {10.1038/s41467-024-50194-3},
  publisher = {Springer Nature}
}
bioRxiv · 2026.03.13.711464 · 2026
OmicClaw: executable and reproducible natural-language multi-omics analysis over the unified OmicVerse ecosystem
Zehua Zeng, Xuehai Wang, Zhi Luo, Yawen Zheng, Lei Hu, Cencan Xing, Hongwu Du
DOI: 10.64898/2026.03.13.711464
View Paper