LLMBDC: LLM-Based Clustering of Gene Ontology Terms into Biological Domains

LLMBDC is a training-free framework that uses zero-shot LLM semantic reasoning and confidence scoring to group redundant Gene Ontology enrichment terms into interpretable biological domains. The authors report evaluations on Alzheimer’s disease and Fragile X syndrome against six baselines. Compared with REVIGO, ARI reportedly increased from 9.7% to 73.3% for Alzheimer’s disease and from 15.7% to 66.6% for Fragile X syndrome, while NMI also improved. The abstract does not provide implementation details, ablations, model identities, or independent verification, and the supplied publication date appears future-dated.