Decoding Cellular Plasticity through Deep Learning of Multi-Omic Gene Regulatory Networks

Explore a cutting-edge research presentation examining how deep learning can decode cellular plasticity in glioblastoma through multi-omic gene regulatory network analysis. Discover the development of scDORI, a scalable deep-learning framework that infers enhancer-driven gene regulatory networks at single-cell resolution using integrated single-nucleus RNA and chromatin accessibility profiles from over one million cells across primary IDH-wildtype glioblastomas. Learn about the structured hierarchy of glioblastoma cell states governed by distinct regulatory programs, with particular focus on how epigenetic plasticity enables or constrains cellular transitions between phenotypic states. Understand the identification of neuronal-like tumor cells as a low plasticity state that employs active repression mechanisms, contrasting with more permissive progenitor-like and astrocytic states. Examine the role of MYT1L as a neuronal-like state-specific repressor that silences master transcription factors of alternative states, and review experimental evidence showing how MYT1L gain-of-function reduces chromatin accessibility, induces neuronal-like identity, and restricts proliferation and invasion in patient-derived glioblastoma cells. Gain insights into the epigenetic architecture and transcriptional master regulators that shape glioblastoma state trajectories, and explore the therapeutic potential of targeting safeguard repressors like MYT1L to constrain malignant plasticity in one of the most lethal human cancers.