Researchers from Columbia University and the Chan Zuckerberg Initiative have developed GREmLN, a graph-based artificial intelligence model designed to capture complex long-range gene regulatory interactions in cells by integrating gene network information. Unlike prior language-based models which rely on sequential logic, GREmLN models gene-gene relationships without sequential assumptions, enabling improved predictive accuracy for cell state transitions in health and disease. GREmLN outperforms existing scRNA foundation models in cell type annotation and understanding gene regulatory graph structures. This advancement supports development of virtual cell models that can simulate cellular behaviors and responses for biology and medical research.