Yale Researchers Develop AI Tool to Precisely Time Gene Activation in Brain Development

ChronODE method offers precision in timing gene therapy treatments

Yale UniversityAug 20 2025 A Yale research team has created a new computer tool that can pinpoint when exactly genes turn on and off over time during brain development - a finding that may one day help doctors identify the optimal window to deploy gene therapy treatments. Dubbed "chronODE," the tool uses math and machine learning to model how gene activity and chromatin (the DNA and protein mix that forms chromosomes) patterns change over time. The tool may offer a variety of applications in disease modelling and basic genomic research and perhaps lead to future therapeutic uses. Basically, we have an equation that can determine the precise moment of gene activation, which may dictate important steps such as the transition from one developmental or disease stage to another. Consequently, this may represent a potential way to identify, in the future, critical points for therapeutic intervention." Mor Frank, postdoctoral associate in the Department of Biophysics and Biochemistry in Yale's Faculty of Arts and Sciences (FAS) and study co-author Results of the study were published August 19 in the journal Nature Communications. For the study, the research team wanted to determine not just when genes activate, but how their activation changes over the course of brain development. Genes activate at different points in cell development, but mapping gene development has been difficult. And past studies have focused on isolated moments in time, not on how gene expression evolves over time. In this case, the researchers used a logistic equation (a mathematical equation useful for modelling dynamic processes) to measure when and how rapidly genes turn on and off in developing mouse brains. They found that most genes follow simple and gradual activation patterns, and that genes can be grouped into subtypes, including accelerators that speed up during late stages of development; switchers that speed up and then slow down; and decelerators that just slow down. Researchers then developed an AI model to predict gene expression over time based on changes in nearby chromatin. The model worked well, especially for genes with a more complex regulation, and the entire procedure established the chronODE method. They found that most genes follow predictable developmental patterns, which are dictated by their role in a cell and determine how quickly they reach maximum influence on the cell. "In a situation where you're treating genetic disease, you'd want to shut down the gene before it reaches its full potential, after which it's too late," said co-author Beatrice Borsari, who is also a postdoctoral associate in biophysics and biochemistry. "Our equation will tell you exactly the switching point - or the point of no return after which the drug will not have the same effect on the gene's expression," Borsari said. "There are many cases where it's not just important to characterize the developmental direction you go, but also how fast you reach a certain point, and that's what this model is allowing us to do for the first time," added Mark Gerstein, the Albert L. Williams Professor of Biomedical Informatics at Yale School of Medicine and a professor of molecular biophysics and biochemistry, computer science, and of statistics and data science in FAS, and the study's lead author. Borsari and Frank underscore that the potential applications in the pharmacokinetic area are major. Researchers called their new method "chronODE," a name that merges the concept of time (Chronos is the god of time in Greek mythology) with the mathematical framework of ordinary differential equations (ODEs.) "We analyze time-series biological data using the logistic ODE," Borsari said. "In a sense, the name captures the multidisciplinary nature of our research. We work where biology meets the beauty of math. We use mathematical models to describe and predict complex biological phenomena - in our case, temporal patterns in genomic data." Borsari is a computational biologist with expertise in genetics and bioinformatics, while Frank is a biomedical engineer with a strong foundation in machine learning and mathematics. "Our diverse skills create a highly synergistic collaboration, and we learn a lot from each other," Borsari said. Other study authors include research associates Eve S. Wattenberg, Ke Xu, Susanna X. Liu, and Xuezhu Yu. Yale University Journal reference: Borsari, B., et al. (2025). The chronODE framework for modelling multi-omic time series with ordinary differential equations and machine learning. Nature Communications. doi.org/10.1038/s41467-025-61921-9.

AI Tool Shows Exactly When Genes Turn On and Off - Neuroscience News

Summary: Researchers have developed an AI-powered tool called chronODE that models how genes turn on and off during brain development. By combining mathematics, machine learning, and genomic data, the method identifies exact "switching points" that determine when genes reach maximum activity. These findings reveal that most genes follow predictable activation patterns and can be classified into subtypes such as accelerators, switchers, and decelerators. The approach could eventually allow doctors to time gene therapies or drug interventions at the most effective moment. A Yale research team has created a new computer tool that can pinpoint when exactly genes turn on and off over time during brain development -- a finding that may one day help doctors identify the optimal window to deploy gene therapy treatments. Dubbed "chronODE," the tool uses math and machine learning to model how gene activity and chromatin (the DNA and protein mix that forms chromosomes) patterns change over time. The tool may offer a variety of applications in disease modeling and basic genomic research and perhaps lead to future therapeutic uses. "Basically, we have an equation that can determine the precise moment of gene activation, which may dictate important steps such as the transition from one developmental or disease stage to another," said Mor Frank, a postdoctoral associate in the Department of Biophysics and Biochemistry in Yale's Faculty of Arts and Sciences (FAS) and study co-author. "Consequently, this may represent a potential way to identify, in the future, critical points for therapeutic intervention." Results of the study were published August 19 in the journal Nature Communications. For the study, the research team wanted to determine not just when genes activate, but how their activation changes over the course of brain development. Genes activate at different points in cell development, but mapping gene development has been difficult. And past studies have focused on isolated moments in time, not on how gene expression evolves over time. In this case, the researchers used a logistic equation (a mathematical equation useful for modeling dynamic processes) to measure when and how rapidly genes turn on and off in developing mouse brains. They found that most genes follow simple and gradual activation patterns, and that genes can be grouped into subtypes, including accelerators that speed up during late stages of development; switchers that speed up and then slow down; and decelerators that just slow down. Researchers then developed an AI model to predict gene expression over time based on changes in nearby chromatin. The model worked well, especially for genes with a more complex regulation, and the entire procedure established the chronODE method. They found that most genes follow predictable developmental patterns, which are dictated by their role in a cell and determine how quickly they reach maximum influence on the cell. "In a situation where you're treating genetic disease, you'd want to shut down the gene before it reaches its full potential, after which it's too late," said co-author Beatrice Borsari, who is also a postdoctoral associate in biophysics and biochemistry. "Our equation will tell you exactly the switching point -- or the point of no return after which the drug will not have the same effect on the gene's expression," Borsari said. "There are many cases where it's not just important to characterize the developmental direction you go, but also how fast you reach a certain point, and that's what this model is allowing us to do for the first time," added Mark Gerstein, the Albert L. Williams Professor of Biomedical Informatics at Yale School of Medicine and a professor of molecular biophysics and biochemistry, computer science, and of statistics and data science in FAS, and the study's lead author. Borsari and Frank underscore that the potential applications in the pharmacokinetic area are major. Researchers called their new method "chronODE," a name that merges the concept of time (Chronos is the god of time in Greek mythology) with the mathematical framework of ordinary differential equations (ODEs.) "We analyze time-series biological data using the logistic ODE," Borsari said. "In a sense, the name captures the multidisciplinary nature of our research. We work where biology meets the beauty of math. We use mathematical models to describe and predict complex biological phenomena -- in our case, temporal patterns in genomic data." Borsari is a computational biologist with expertise in genetics and bioinformatics, while Frank is a biomedical engineer with a strong foundation in machine learning and mathematics. "Our diverse skills create a highly synergistic collaboration, and we learn a lot from each other," Borsari said. Other study authors include research associates Eve S. Wattenberg, Ke Xu, Susanna X. Liu, and Xuezhu Yu. Author: Bess Connolly Source: Yale Contact: Bess Connolly - Yale Image: The image is credited to Neuroscience News Original Research: Open access. "The chronODE framework for modelling multi-omic time series with ordinary differential equations and machine learning" by Eve S. Wattenberg et al. Nature Communications Abstract The chronODE framework for modelling multi-omic time series with ordinary differential equations and machine learning Many genome-wide studies capture isolated moments in cell differentiation or organismal development. Conversely, longitudinal studies provide a more direct way to study these kinetic processes. Here, we present an approach for modeling gene-expression and chromatin kinetics from such studies: chronODE, an interpretable framework based on ordinary differential equations. chronODE incorporates two parameters that capture biophysical constraints governing the initial cooperativity and later saturation in gene expression. These parameters group genes into three major kinetic patterns: accelerators, switchers, and decelerators. Applying chronODE to bulk and single-cell time-series data from mouse brain development reveals that most genes (~87%) follow simple logistic kinetics. Among them, genes with rapid acceleration and high saturation values are rare, highlighting biochemical limitations that prevent cells from attaining both simultaneously. Early- and late-emerging cell types display distinct kinetic patterns, with essential genes ramping up faster. Extending chronODE to chromatin, we find that genes regulated by both enhancer and silencer cis-regulatory elements are enriched in brain-specific functions. Finally, we develop a bidirectional recurrent neural network to predict changes in gene expression from corresponding chromatin changes, successfully capturing the cumulative effect of multiple regulatory elements. Overall, our framework allows investigation of the kinetics of gene regulation in diverse biological systems.