Using the power of AI to make genetic diagnosis easier

HOUSTON – (May 21, 2026) – Researchers at the Jan and Dan Duncan Neurological Research Institute at Texas Children’s Hospital (Duncan NRI) and Baylor College of Medicine have developed a new computational tool, MARRVEL-MCP, designed to help clinicians and scientists move more efficiently toward genetic diagnoses by analyzing and interpreting vast amounts of genetic and biological data using everyday language. The study appears in the American Journal of Human Genetics.

“Rare genetic diseases are often caused by small changes in a person’s DNA, but determining which of those changes are truly disease-causing can be incredibly complex,” said co-corresponding author Dr. Hyun-Hwan Jeong, principal investigator at the Duncan NRI. “This process requires evaluating large datasets, which can be time-consuming and challenging even for experts.”

“To reach a genetic diagnosis, clinicians and researchers must integrate information from multiple biological databases, each with its own structure and requirements,” said co-corresponding author Dr. Zhandong Liu, chief of computational sciences at Texas Children’s Hospital, principal investigator at the Duncan NRI and Associate Professor at Baylor College of Medicine. “MARRVEL-MCP streamlines this process, enabling faster and more accessible interpretation.”

From MARRVEL to MARRVEL-MCP
MARRVEL-MCP builds on MARRVEL (Model organism Aggregated Resources for Rare Variant ExpLoration), a widely used resource that integrates genomic, functional, and model-organism data to support the interpretation of genetic variants. The platform is used globally by tens of thousands of researchers each year.

While MARRVEL consolidated critical data sources, it still required technical expertise to interpret results. MARRVEL-MCP advances this capability by allowing users to interact with the system using plain language. Instead of navigating multiple databases manually, users can ask questions such as, “Is this BRCA1 mutation linked to cancer?”

In seconds, the system identifies relevant genetic features, queries multiple curated databases, and synthesizes results into a clear, evidence-based response. The tool leverages artificial intelligence, including large language models, to automate multi-step analyses across areas such as disease associations, genetic variation, gene expression, and scientific literature.

“What excites us most is the ability to make advanced computational tools more accessible,” Jeong said. “By combining curated biomedical resources with AI, we are helping enable broader use of genetic analysis in both research and clinical settings.”

“We’ve made MARRVEL-MCP available as an open resource,” Liu said. “Through a publicly accessible interface, users can explore the system without requiring specialized installations. This reflects our commitment at Texas Children’s to advancing collaborative, innovative approaches to rare disease research.”

Collaboration and support
First author Zachary Everton, along with collaborators across Texas Children’s Hospital and affiliated institutions, contributed to this work.

This research was supported by multiple organizations, including the Cancer Prevention and Research Institute of Texas, the Chan Zuckerberg Initiative, the National Institutes of Health, and the Duncan NRI, among others.