GRN-Guided Simulation of Single-Cell RNA-Seq and Primer on Causal GANs

Explore a comprehensive lecture from the Broad Institute's Models, Inference and Algorithms series featuring two speakers. Begin with Yazdan Zinati from McGill University, who provides a primer on Causal Generative Adversarial Networks (GANs), offering theoretical background on both traditional and causal GANs. Then, delve into Professor Amin Emad's presentation on GRouNdGAN, an innovative gene regulatory network (GRN)-guided causal implicit generative model for simulating single-cell RNA-seq data. Learn how GRouNdGAN captures non-linear transcription factor-gene dependencies, preserves various cellular characteristics, and outperforms existing simulators in generating realistic cells. Discover its applications in performing in-silico perturbation experiments and benchmarking GRN inference methods, bridging the gap between simulated and biological data benchmarks. Gain insights into how GRouNdGAN provides gold standard ground truth GRNs and realistic cells corresponding to biological systems of interest, making it a valuable tool for single-cell RNA-seq analysis.