New study introduces advanced method to better understand how cells change over time

A recent study published in Nature Communications introduces a new computational approach that helps scientists better understand how cells develop and change over time by examining DNA methylation—a key chemical process that regulates gene activity.

DNA methylation acts like a molecular switch, turning genes on or off without changing the underlying DNA sequence. In recent years, advances in single-cell sequencing have allowed researchers to study methylation patterns at the level of individual cells. This is especially important because cells in the same tissue can behave very differently depending on their stage of development or function. However, analyzing how these methylation patterns evolve over time—particularly during processes like embryonic development—has remained a major challenge.

To address this, co-authors and researchers at UTHealth Houston School of Public Health, Hao Feng, PhD, associate professor in the Department of Biostatistics and Data Sciences, and Wen Tang, MS, helped develop a new method called “MIST” (methylation inference for single cells along trajectories). This tool uses a hierarchical Bayesian framework—a statistical modeling approach that can account for variability and uncertainty in complex biological data. The method tracks cells along a developmental timeline (often referred to as “pseudotime”) and identifies where meaningful changes in DNA methylation occur.

What makes MIST particularly valuable is its ability to detect subtle, stage-specific changes in methylation across individual cells. In testing, the method outperformed existing approaches in identifying genes with significant methylation changes over time.

“By tracking those changes one cell at a timeover pseudotime, our work could help researchers find earlier warning signs of disease, discover better biomarkers, and better understand the biological steps that lead from healthy cells to diseased ones. Over time, that kind of knowledge can support earlier diagnosis and more precise treatment strategies,” said Feng.

The researchers applied MIST to real-world datasets, including mouse embryo development and the developing human brain. In both cases, the tool successfully identified key regulatory genes and methylation patterns that align with major developmental transitions—offering new insight into how cells specialize and form tissuescommit to specialized cell types.

This work represents an important step forward in epigenetics and single-cell analysis. By providing a more precise way to map how gene regulation evolves during development, MIST could help scientists better understand diseases linked to epigenetic changes, including cancer and neurological disorders. It also opens new opportunities for studying how environmental factors influence gene activity over time. “A key strength of our research is that it offers a fundamental analytical approach that can be used to study many different diseases,” said Feng.

Publications such as this one are one small step towards solving some of the greatest health challenges. Feng shared, “I’m excited by the possibility of making very complex biological data more understandable, so we can see patterns that were hidden before. I’m also motivated by the fact that building a computational method does not just help one lab, it can help many other researchers ask better questions and make new discoveries.”

Ultimately, tools like mist are helping transform large, complex biological datasets into actionable insights—bringing researchers closer to decoding the dynamic processes that shape human health and disease.

Additional authors on the study include Daoyu Duan, Case Western Reserve; Wenjing Ma, University of Michigan Ann Arbor; Hao Wu, Emory University; and Liangliang Zhang, Case Western Reserve.