Biologically Guided Variational Inference for Interpretable Multimodal Single-Cell Integration

NetworkVI is a sparse deep generative model designed for the integration and interpretation of multimodal single-cell data. NetworkVI models gene-gene interactions inferred from topological associated domains and utilizes the structure of gene ontology (GO) to aggregate gene embeddings to cell embeddings, enhancing the interpretability at the gene and GO-term level. NetworkVI can be used for modality imputation, reference-to-query mapping and aids in identifying modality- and cell type-specific signatures via interpretability. NetworkVI will support researchers in interpreting cellular disease mechanisms, guiding biomarker discovery, and ultimately aiding the development of targeted therapies in large-scale single-cell multimodal atlases.

Check out the API and Tutorials section for further information.

If you use NetworkVI, please consider citing:

Arnoldt, L., Upmeier zu Belzen, J., Herrmann, L., Nguyen, K., Theis, F.J., Wild, B. , Eils, R., “Biologically Guided Variational Inference for Interpretable Multimodal Single-Cell Integration”, bioRxiv, June 2025.

Installation

1. Install the latest release of NetworkVI from PyPi:

pip install networkvi

2. Install the latest development version:

pip install git+https://github.com/LArnoldt/networkvi.git@main

Contents

• API
  • Model
    • networkvi.model.NETWORKVI
      • NETWORKVI
        • NETWORKVI.__init__()
  • Layers
    • networkvi.nn.GeneLayers
      • GeneLayers
        • GeneLayers.__init__()
    • networkvi.nn.GOLayers
      • GOLayers
        • GOLayers.__init__()
• Tutorials
  • Integration
    • Paired integration and query-to-reference mapping
      • Data loading
      • Data setup
      • Model setup and training
        • NetworkVI
      • Query-to-reference mapping
    • Mosaic integration
      • Data loading
      • Data setup
      • Model setup and training
        • NetworkVI
  • Interpretability
    • Interpretability: Inferring importances for GO and genes
      • Data loading
      • Data setup
      • Model setup and training
        • NetworkVI
      • Interpretability
    • Interpretability: Inferring GO-specific covariate attention values
      • Data loading
      • Data setup
      • Model setup and training
        • NetworkVI
      • Interpretability
• Changelog
• References