This research evaluates how large language models can revolutionize the identification of causal genes for rare genetic diseases, addressing a critical diagnostic challenge in medicine.

• Benchmarked various LLMs for gene prioritization effectiveness
• Implemented multi-agent approaches with Human Phenotype Ontology (HPO) classification
• Developed improvement strategies to enhance LLM performance in clinical settings
• Demonstrated potential to overcome diagnostic challenges posed by limited patient data and genetic diversity

This work matters because rare diseases affect millions worldwide but remain difficult to diagnose due to their genetic complexity. LLM-assisted gene prioritization could significantly reduce diagnostic odysseys for patients, enabling earlier interventions and treatments.

Survey and Improvement Strategies for Gene Prioritization with Large Language Models