HBM Holdings Limited announced that its wholly owned subsidiary, Nona Biosciences, has developed HCAbLM, described as the world’s first language model trained exclusively on fully human heavy-chain-only antibodies (HCAbs). The research, titled “A foundation model learns the sequence and functional grammar of fully human heavy-chain-only antibodies,” is now available on bioRxiv.
HCAbLM was built on a repertoire of 31.80 million fully human HCAb sequences derived from 73 independently immunized HCAb mice, enabling the model to learn a distinct “sequence grammar” that conventional antibody language models do not capture. According to HBM, HCAbLM’s representations correlate with experimentally measured attributes—size-exclusion chromatography (SEC) purity, hydrophobic interaction chromatography (HIC) behavior, and thermal stability—supporting its potential role in evaluating and optimizing fully human HCAbs beyond sequence analysis alone.
The development aligns with Nona Biosciences’ broader strategy to integrate Harbour Mice® transgenic technology, AI-driven discovery via Hu-mAtrIx™, and end-to-end preclinical services. Management positions HCAbLM as a foundational tool to accelerate data-driven antibody design and developability assessments across modalities such as bispecific antibodies, CAR-T therapies, antibody-drug conjugates, and mRNA-based therapeutics.
HBM reiterated that no assurance can be given on the successful development or commercialization of any drug candidate arising from this platform and advised shareholders and potential investors to exercise caution when dealing in the company’s shares.
Comments