Mayo Clinic researchers have unveiled an AI tool named UNISOM, capable of detecting early blood cell mutations associated with higher risks of leukemia and heart disease.
In a significant advancement for early disease detection, Mayo Clinic researchers have developed an artificial intelligence tool designed to identify early mutations in blood cells. These mutations can significantly increase the risk of leukemia and heart disease in older adults.
The tool, named UNISOM — short for Unified Somatic Calling and Machine learning — is detailed in a study published in the journal Genomics, Proteomics & Bioinformatics.
The AI technology has shown impressive potential in spotting early signs of clonal hematopoiesis of indeterminate potential (CHIP), a condition that often goes unnoticed but is linked with heightened risks of serious health issues.
CHIP originates in the bone marrow, where stem cells produce blood cells vital for various bodily functions.
Occasionally, these stem cells mutate abnormally and multiply, forming clusters of mutated cells. This can happen silently over time, significantly raising the risk of developing leukemia by more than tenfold and increasing the risk of heart disease up to fourfold, even in otherwise healthy adults.
“Detecting disease at its earliest molecular roots is one of the most meaningful advances we can make in medicine,” co-senior author Eric Klee, the Everett J. and Jane M. Hauck Midwest Associate Director of Research and Innovation, said in a news release.
UNISOM was developed by Shulan Tian, under Klee’s leadership, and helps clinicians identify CHIP-related mutations within standard genetic datasets.
Prior to this innovation, such precise detection required complex and advanced sequencing methods.
The tool detected nearly 80% of CHIP mutations using whole-exome sequencing, which targets the DNA regions coding for proteins.
Moreover, when tested on whole-genome sequencing data from the Mayo Clinic Biobank, UNISOM successfully identified early CHIP signs, including mutations present in fewer than 5% of blood cells — changes often missed by conventional techniques.
“We’re engineering a path from genomic discovery to clinical decision-making,” added Tian, a bioinformatician at Mayo Clinic and co-senior author of the study. “It’s rewarding to help bring these discoveries closer to clinical care, where they can inform decisions and support more precise treatment.”
The Mayo Clinic team intends to apply UNISOM to larger and more varied datasets, aiming to refine the tool further and expand its practical applications in clinical environments.
Source: Mayo Clinic