Core Viewpoint - The article discusses the development of CellNavi, an AI framework designed to map and manipulate cell fate, addressing a long-standing challenge in cell biology and potentially transforming regenerative medicine, cancer research, and drug development [2][3][12]. Group 1: Introduction to Cell Fate and Plasticity - The concept of "developmental landscape" introduced by Conrad Waddington illustrates that cell fate is not unidirectional and can be altered, leading to the exploration of how to control this fate [2]. - Advances in stem cell research and reprogramming experiments have shown that cells can be pushed back to higher states or redirected to new paths, raising the question of whether we can map and manipulate this fate [2]. Group 2: The Need for Systematic Tools - Traditional methods in cell biology rely on large-scale experiments to identify candidate genes, which is inefficient and may overlook critical factors [3]. - There is a pressing need for systematic tools that can accurately locate cell states and predict regulatory measures to guide them toward target states [3]. Group 3: CellNavi Framework - CellNavi utilizes deep learning to capture low-dimensional manifolds of cell states, acting as a "navigation tool" for researchers to identify key factors driving cell state transitions [3][5]. - The framework integrates large-scale single-cell transcriptomic sequencing and CRISPR perturbation data to create a system that maps and navigates cell states [5]. Group 4: Functionality of CellNavi - CellNavi predicts driving genes for transitions between given starting and target cell states, ranking them based on their influence [7][8]. - It combines static single-cell data with dynamic predictions, allowing for the identification of genes that can induce significant changes in cell states [8]. Group 5: Experimental Validation and Performance - Experimental results indicate that CellNavi outperforms traditional algorithms in various benchmark tasks, demonstrating its ability to identify key factors even when gene expression levels do not show significant changes [9][11]. - The framework has shown potential in drug mechanism elucidation, revealing different effects of HDAC inhibitors on downstream pathways despite targeting the same molecular entities [11]. Group 6: Future Directions and Research Paradigms - CellNavi represents a new research paradigm, enabling the extraction of systematic knowledge from complex data and transforming abstract developmental landscapes into actionable navigation maps [12]. - Future research may focus on integrating genomics, spatial omics, and epigenetics to enhance the model's accuracy and interpretability, establishing a tighter feedback loop with experimental validation [13].