AI Helps To Interpret Genetic Variants, Progressing Rare Disease Diagnosis

A new computational tool called MARRVEL-MCP helps researchers move toward genetic diagnoses more efficiently by analyzing and interpreting vast amounts of genetic and biological information using everyday language. The study, conducted by researchers at Baylor College of Medicine and Texas Children’s Hospital, appeared in the American Journal of Human Genetics.

“Rare genetic diseases are often caused by small changes in a person’s DNA. However, not all genetic changes linked to a condition may play a role in the disease,” said co-corresponding author Dr. Hyun-Hwan Jeong, assistant professor of pediatrics – neurology at Baylor and investigator at the Duncan Neurological Research Institute at Texas Children’s. “Some changes may contribute to disease, while others may not. Identifying whether a particular genetic change or variant is harmful or an innocent bystander is crucial for diagnosing these conditions, but the process requires sifting through large amounts of data, a complex and time-consuming task.”

“To reach a genetic diagnosis, doctors and researchers must gather information from many different biological databases, each with its own format and rules, and then carefully piece together the evidence. Even for experts, this can take hours for a single case,” said co-corresponding author Dr. Zhandong Liu, associate professor of pediatrics – neurology at Baylor. Liu also is chief of computational sciences at Texas Children’s. 

The current study introduces MARRVEL-MCP, a new computational tool that is designed to make this process faster and more accessible, especially for non-experts. It combines artificial intelligence, specifically large language models (LLMs) like ChatGPT and Gemini, with a structured set of biological databases to help interpret genetic variants using layman’s terms.

From MARRVEL to MARRVEL-MCP

The team previously had developed MARRVEL (Model organism Aggregated Resources for Rare Variant ExpLoration), a computational approach that allows researchers to comb in a matter of minutes through large genetic and biological databases all at once to search for information regarding gene variants. MARRVEL has been well received by the scientific community, recording more than 43,000 users worldwide in 2025 alone.

MARRVEL brings together genomic, functional and model-organism databases into a unified platform. These sources contain different types of information that need to be considered to determine whether a genetic variant causes a disease. For instance, how common a variant is in the population, whether it has been linked to disease before, predictions about whether it damages a gene, information from lab experiments and model organisms and scientific articles discussing similar cases.

“However, MARRVEL requires precisely formatted inputs and produces comprehensive but complex outputs that demand substantial manual interpretation,” Jeong said. “This poses barriers that limit its accessibility and efficiency for many users as it assumes they can interpret heterogeneous outputs and synthesize evidence across sections, which requires substantial expertise.”

MARRVEL-Model Context Protocol (MCP) changes how this process works. Instead of requiring users to learn technical formats and manually navigate databases, it allows them to ask questions in plain language, such as, “Is this BRCA1 mutation linked to cancer?”

In a matter of seconds, MARRVEL-MCP automatically identifies key pieces of information (like gene names or mutations), converts them into the formats required by databases, queries multiple data sources in the correct order and combines the results into a clear, evidence-based answer. MARRVEL-MCP covers areas like disease associations, genetic variation, gene expression and scientific literature and enables LLMs to autonomously compose and execute multi-step analytical workflows from simple language queries.

“What excites me most is that MARRVEL-MCP shows we do not always need the largest frontier AI models to make meaningful progress in biomedical research,” Jeong said. “By giving smaller models access to the right curated tools and structured context, we can make them smarter for specialized tasks. For example, gpt-oss-20b, a model that can be installed locally, improved to 94% with MARRVEL-MCP from 41% accuracy without MARRVEL-MCP. This suggests a path toward more accessible and cost-effective AI for rare disease research.”

Reference: Everton Z, Botas J, Kim SY, et al. MARRVEL-MCP: An agentic interface for Mendelian disease discovery via tool-augmented context engineering. Am J Hum Genet. 2026. doi: 10.1016/j.ajhg.2026.04.012

This article has been republished from the following materials. Note: material may have been edited for length and content. For further information, please contact the cited source. Our press release publishing policy can be accessed here.