In a study published in Cell titled, “Deep learning of functional perturbations from condensate morphology,” researchers at Princeton University have applied AI to understand how drugs affect the dynamics of key structures within the cell. The work introduces a tool that can map morphology to functional outcomes and shed light on markers of health. 

The authors examined the changes in shape of biomolecular condensates, tiny droplets in cells that drive transcription and other gene regulation processes linked to disease, including Alzheimer’s, ALS and cancer. The findings support a robust system for monitoring and evaluating cellular responses to drugs at a single-cell level. 

“The central problem in biology is how do you get emergent structure from individual molecular interactions,” said Cliff Brangwynne, PhD, professor of chemical and biological engineering at Princeton and corresponding author of the study. “The key innovation here was to develop a way to learn from the images and classify the patterns that are emergent.” 

The team used an advanced microscope to image nucleolar morphology changes in hundreds of human cells under a range of drug-controlled conditions. Machine learning tools sorted the images into four basic categories based on the shape of the nucleolus, uncovering “cap” and “necklace” shapes linked to cellular stress responses.

The authors ran a panel of drugs to examine the effect on nucleolar formation and measured changes in the condensate’s development. Varying concentrations caused different degrees of change in both caps and necklaces.  

Two known anti-cancer drugs caused caps, while a third drug, called topotecan, triggered a new nucleolus morphology that the researchers labeled “flower.” While topotecan inhibits TOP1, an key enzyme during DNA replication, loss of TOP1 induced the flower shape and uncovered the enzyme’s role in maintaining nucleolar organization by regulating RNA processing. 

“No one’s seen this flower morphology before,” said Brangwynne. “The network flagged it as not fitting neatly into the other three categories.” 

The team also tested their neural network on other condensates related to RNA processes, observing similar dose-and-response results for drugs specific to nuclear speckles, a hub for messenger RNA activity, and condensates from respiratory syncytial virus. 

This finding underscores the value of analyzing morphological changes. “You could be missing other important features,” said Anita Donlic, PhD, postdoctoral researcher and first author of the study. “Things that could tell you there’s new biology.” 

The post AI Predicts Gene Regulation for Drug Discovery Using Condensate Morphology appeared first on GEN – Genetic Engineering and Biotechnology News.