Welcome to CellCap’s documentation!
CellCap is a variational autoencoder to model correspondence between cellular identity and perturbation response.
- Introduction
- Installation
- Apply CellCap to simulation data
- Load packages
- Prepare data for ready-to-use
- Learning good basal state is critical for establishing correspondence between cellular identity and perturbation response
- 1. Train CellCap without adversarial loss
- 2. A small penalty on adversarial learning
- 3. Keep the penalty on adversarial learning large
- A structured outcome requires tuning penalty on automatic relevance determination
- The right solution for this simulation data
- Apply CellCap to human monocytes data
- Citation