scPreGAN

The software tool 'scPreGAN' is designed to predict the cellular responses of individual cells to perturbations using single-cell RNA sequencing data. It addresses the challenge of understanding how different cell types respond to perturbations, such as drug treatments, without the need to collect cells after the perturbation. This prediction approach can save time and resources.

Key features of scPreGAN:
1. Deep Generative Model: scPreGAN employs a deep generative model that integrates an autoencoder and a generative adversarial network (GAN). The autoencoder extracts common information from both unperturbed and perturbed data, while the GAN predicts the perturbed data.
2. Improved Prediction Accuracy: The tool has been tested on three real datasets and has demonstrated superior prediction accuracy compared to three state-of-the-art methods. It is capable of capturing the complex distribution of cell expression and generating prediction data that closely matches real data in terms of expression abundance.
scPreGAN provides a solution for predicting how individual cells respond to perturbations, which is especially useful when collecting perturbed cells is challenging or impractical.