HBM HOLDINGS-B(02142): Nona Biosciences Successfully Develops the World's First Language Model Based on Fully Human Heavy-Chain-Only Antibodies

HBM HOLDINGS-B(02142) announced that its wholly-owned subsidiary Nona Biosciences (Nona) has successfully developed HCAbLM, the world's first language model trained on fully human heavy-chain-only antibodies (HCAbs). The related research paper, "A foundation model learns the sequence and functional grammar of fully human heavy-chain-only antibodies," has been published on bioRxiv. HCAbLM was built on a large-scale database comprising 31.8 million fully human HCAb sequences from 73 independently immunized HCAb transgenic mice, aiming to capture the unique sequence features of fully human HCAbs. Research shows that HCAbLM has learned a unique sequence "grammar" associated with the heavy-chain-only antibody format, revealing sequence patterns and their interrelationships that conventional antibody language models cannot adequately capture. More importantly, the representations learned by the model also demonstrate transferability to experimentally measured antibody properties, including size exclusion chromatography (SEC) purity, hydrophobic interaction chromatography (HIC) behavior, and thermal stability, highlighting HCAbLM's potential in the evaluation and optimization of fully human HCAbs, going beyond mere sequence analysis. The development of HCAbLM marks a significant advance by Nona in applying artificial intelligence (AI) to antibody discovery and development. By combining large-scale antibody libraries with AI-driven modeling, Nona is laying the foundation for more efficient, data-driven antibody design and developability assessment.