Nona Biosciences Successfully Develops World's First Language Model Trained on Fully Human Heavy-Chain-Only Antibodies
Nona Biosciences developed HCAbLM, the first language model trained on fully human heavy-chain-only antibodies (HCAbs). The model used 31.8 million HCAb sequences from 73 immunized mice. HCAbLM addresses a critical bottleneck in AI antibody development: developability, which includes manufacturability, aggregation, and stability. The model demonstrated superior performance in predicting developability metrics compared to larger general-purpose protein models like Meta's ESM-6B.
Nona Biosciences' HCAbLM model is a specific advance in biotech AI. It targets the 'developability' of heavy-chain-only antibodies, not just their binding. This focus on practical drug characteristics like stability and manufacturability is crucial. The model, with 366 million parameters, outperformed Meta's 6-billion-parameter ESM-6B on key metrics. This shows specialized models can beat general-purpose ones in niche applications.
For Asia's biopharma sector, this development points to accelerated drug discovery. Companies can use AI to screen drug candidates more efficiently. Nona Biosciences, a subsidiary of HBM Holding Limited (HKEX:02142), strengthens its position. This could reduce R&D costs and speed up time to market for new biotherapeutics across the region.
The test for Nona Biosciences is integration. HCAbLM must seamlessly merge with existing drug discovery platforms. Its real impact will be measured by the number of successful drug candidates it helps advance. This will determine if it moves beyond a research success to a commercial advantage.
Related reading
6 stories
China’s rocket boom turns Hainan into a space hub. Can launches fuel wider growth?
Anthropic merges Claude chat and Cowork in one interface
Threads’ new features let podcasters promote shows and reach listeners
China’s Z.ai raises revenue target 25% after US$5 billion cash injection
SK Hynix reportedly in talks with Intel to build memory chips in US
