Researchers at the University of Florida, led by Dr. Kiley Graim, have developed PhyloFrame, an innovative AI tool designed to reduce 'ancestral bias' in genetic research, ultimately enhancing precision medicine for diverse populations.

PhyloFrame introduces Enhanced Allele Frequency (EAF), a novel measure that identifies population-specific genetic variants in healthy tissue, thereby improving disease predictions.

The research team is addressing the critical issue of genetic data predominantly representing a single ancestral group, which limits advancements in disease treatment and prevention.

To process the vast amounts of genomic data required for this research, the team utilized UF's HiPerGator supercomputer, analyzing approximately 3 billion base pairs of DNA per individual.

Despite the advancements made with PhyloFrame, the researchers stress the ongoing need for more diverse genetic databases to support equitable precision medicine and effective machine learning modeling.

Dr. Graim pointed out that a staggering 97% of sequenced genetic samples originate from individuals of European ancestry, highlighting a significant bias that affects the accuracy of genetic research.

The effectiveness of PhyloFrame was detailed in a study published on March 10, 2025, in Nature Communications, with support from the National Institutes of Health.

Funding for the project also came from the UF College of Medicine's AI2 Datathon grant, which aims to leverage AI tools for health advancements.

The researchers emphasized that incorporating diverse training data not only benefits underrepresented populations but also enhances model performance for European groups, preventing overfitting.

PhyloFrame distinguishes itself by maintaining prediction accuracy across various populations, addressing the shortcomings of existing models that primarily rely on data from European ancestry.

The ultimate goal for PhyloFrame is to be implemented in clinical settings, facilitating personalized treatment plans based on genetic profiles while minimizing side effects.

Dr. Graim's interest in addressing ancestral bias was sparked by a conversation with a doctor concerned about the limited relevance of genetic studies for his diverse patient population.