KAIST Develops NUDGE: Single-Stimulus Framework for Precise Cell Fate Redirection

A research team at KAIST in South Korea has developed a computational framework called NUDGE that solves a fundamental problem in cell biology: how to redirect a cell toward a desired state without permanently locking its genes into place. The work, published in September 2026 in the Proceedings of the National Academy of Sciences, offers a new approach to stem cell differentiation, immune modulation, and disease reversal—applications that have long promised breakthroughs in regenerative medicine but have been hampered by the rigidity of existing methods.

The question driving the research is deceptively simple: which genes must be controlled to guide a stem cell into becoming, say, a heart muscle cell, or to calm an overactive immune response? Existing technologies have answered this by locking specific genes permanently in the "on" or "off" position. The problem is that this permanence comes with a cost. Cells lose their plasticity—their ability to adapt and respond to future changes in their environment. The forced state can also produce abnormal cell types that do not occur naturally, creating new problems even as it solves the original one.

Professor Kwang-Hyun Cho's team took a different approach. Rather than forcing a cell into a rigid configuration, they designed a system that nudges the cell's own internal gene networks toward the desired outcome through a single, temporary intervention. The analogy is apt: imagine a blocked water channel. Instead of permanently redirecting the flow, you briefly clear the obstruction and let the water find its natural course. The cell, in this framework, already possesses the capacity to reach the target state; it simply needs a momentary push in the right direction.

NUDGE works by building a computational logic model of how genes inside a cell interact with one another. The system then mathematically decomposes these interactions to identify the minimal set of molecular targets that need temporary activation or inhibition to move the cell from its current state to the desired one. The framework is mathematically guaranteed to find all minimal control combinations that produce the target outcome. For large-scale networks with thousands of molecular interactions, the team developed an efficient approximation method that keeps the approach practical.

The researchers tested NUDGE on three real biological processes. In cardiomyocyte differentiation—the conversion of stem cells into heart muscle cells—the framework reproduced known biological patterns, showing how a single temporary intervention involving the gene MESP1 could guide the process without keeping MESP1 continuously active. In mast cell formation, the team identified GATA2 and GATA1 as key regulators. In macrophage immune cells, they derived strategies for temporary intervention using molecules like IL-4 to shift cells toward an anti-inflammatory state while preserving their ability to switch between different phenotypes.

The performance metrics are substantial. When tested against 552 control problems across 69 large biological networks, NUDGE's approximation method achieved an average intervention error below 0.01 in more than 90 percent of cases. Across all three computational methods tested, NUDGE identified error-free interventions in 83 percent of minimum-sized control problems, compared with 59 percent for a competing method called IBMFA and 53 percent for LDOI. In a separate analysis of 63 published biological network models, the team found that 55 of them showed reduced plasticity or abnormal attractors when permanent control was applied—a finding that underscores the advantage of temporary intervention.

The significance extends beyond the technical achievement. NUDGE is positioned as a general-purpose framework applicable across multiple domains: guiding stem cells into specific cell types for tissue repair, reversing the effects of aging or disease in existing cells, modulating immune responses in autoimmune conditions, and even designing therapies that could push cancer cells back toward normal states. Professor Cho emphasized that the technology is not limited to returning cells to health but functions as a foundational tool for designing control strategies across regenerative medicine, aging research, autoimmune disease, and cancer reversion. The work was supported by multiple grants from South Korea's National Research Foundation and Ministry of Science and ICT, reflecting the government's investment in foundational biomedical research.

The research team included Ferio Brahmana, a master's degree holder, and Corbin Hopper, a PhD student, as co-first authors, with PhD student Woojeong Lee contributing as well. Their findings suggest that the next generation of cell-based therapies may rely not on forcing cells into permanent new configurations but on understanding and leveraging the dynamic systems already present within them.