NLM DIR Seminar Schedule

Abstract:

Gene set analysis plays a central role in genomic discovery, yet conventional approaches often yield results that are difficult to interpret, lack transparency, and fail to account for biological context. As datasets grow in scale and complexity, the gap between statistical output and actionable biological insight continues to widen. In this presentation, I will introduce a new paradigm that integrates large language models into gene set analysis to bridge this gap. I will present three complementary frameworks. GeneAgent mitigates hallucination and enhances explanation robustness, enabling more reliable biological interpretation. Gene-R1 fine-tunes small language models to support privacy-preserving analysis and local deployment, making advanced genomic reasoning accessible in secure environments. cGSA introduces contextual awareness to functional prioritization, improving biological relevance and interpretability.