Yale Team Develops Imaging Technique Linking Aging and Disease

A research team at Yale University has unveiled a groundbreaking imaging technique that reveals crucial connections between aging, disease, and genetic activity in human cells. Utilizing a novel machine learning approach, the researchers discovered that tissue samples viewed under a microscope can uncover genetic variants, assess gene activity, and even estimate an individual’s age.

Ran Meng, the study’s lead author and a postdoctoral researcher in Yale’s Department of Molecular Biophysics and Biochemistry, emphasized the significance of the findings. “Our study shows that ordinary tissue images contain patterns that can reliably predict gene expression and reveal a person’s age—information that was previously hidden to the naked eye,” Meng stated. He noted that the enhanced image quality allows for better links between genetic features and cellular characteristics.

The implications of this research are profound. By improving diagnostic practices through routine pathology slides, the technique may enable early detection of abnormal tissue patterns, which could help predict disease risks. The research findings are published in the Proceedings of the National Academy of Sciences.

Exploring Genetic Connections

Co-author Mark Gerstein, the Albert L. Williams Professor of Biomedical Informatics at Yale School of Medicine, explained the importance of the genotype-phenotype connection in genetics. The genotype is an organism’s genetic makeup, while the phenotype encompasses observable characteristics influenced by both genetics and environmental factors. These can include traits like height and eye color, as well as more complex conditions such as behavioral traits or diseases.

“One of the new research frontiers is ‘multi-modality’—connecting genotype to various types of data that describe phenotype,” Gerstein said. “In this paper, we make an advance in connecting genotype to image features.”

Machine Learning Reveals Hidden Patterns

In their study, the researchers employed machine learning techniques to analyze tissue images from healthy human donors. This analysis enabled them to uncover hidden indicators of aging and gene activity present in human cells. The appearance of cells is intricately linked to both genetics and the aging process.

Using histology slides, genetic information, and RNA data from 838 donors across 12 different tissue types and over 10,000 images, the team developed computer models capable of identifying genetic variants associated with tissue appearance. These models also predicted gene expression—indicating when genes are active—and provided estimates of a person’s age.

Notably, one of the machine learning models demonstrated the ability to predict gene expression based on tissue images, with particularly strong accuracy in samples from the lung, heart, and testis. Another model estimated chronological age from tissue samples, with the skin, tibial nerve, tibial artery, and testis tissue yielding the most accurate predictions due to significant age-related changes.

The study highlighted that the shape, size, and structure of cell nuclei carry substantial biological information. Researchers identified 906 points in the human genome that were closely tied to the appearance of nuclei in various tissues, revealing strong correlations between nuclear shape and gene activity.

The innovative work of the Yale team not only advances scientific understanding but also holds promise for enhancing diagnostic techniques, ultimately improving patient care. As they continue to explore the complex relationships between genetics and observable traits, this research could pave the way for significant developments in the fields of genomics and pathology.