Cambridge – September 18, 2026 -- Nona Biosciences has developed HCAbLM, the first language model trained specifically on fully human heavy-chain-only antibodies (HCAbs), built from a dataset of 31.8 million HCAb sequences drawn from 73 independently immunized transgenic mice. The findings were detailed in a study titled "A foundation model learns the sequence and functional grammar of fully human heavy-chain-only antibodies," published on bioRxiv.
Nona targets the developability bottleneck in AI antibody design
Current AI antibody discovery is limited less by target-binding design than by manufacturability at scale, aggregation risk and stability -- collectively known as developability. HCAbLM was built to learn the underlying sequence grammar that determines whether an HCAb can fold correctly and progress into a viable drug candidate. General-purpose protein models lack this capability because their training data contain almost no HCAb sequences.
A 366-million-parameter model beats a 6-billion-parameter rival on core benchmarks
In public cross-project developability benchmarks, HCAbLM outperformed Meta's 6-billion-parameter ESM-6B, along with established antibody language models IgLM and AbLang, on two key metrics: size exclusion chromatography (SEC) purity and hydrophobic interaction chromatography (HIC) behavior. HCAbLM achieved this with only 366 million parameters, a fraction of ESM-6B's scale.
Model representations correlate with experimentally measured antibody properties
HCAbLM's learned representations transferred to lab-measured properties including SEC purity, HIC behavior and thermal stability, indicating utility beyond raw sequence analysis for evaluating and optimizing fully human HCAbs.
"The development of HCAbLM represents an important step in our efforts to apply AI to specialized antibody formats and unlock the value of large-scale antibody repertoires," said Dr. Di Hong, Chief Executive Officer of Nona Biosciences. She said the model provides a new foundation for AI-enabled antibody discovery and is intended to integrate with Nona's existing technology platforms to accelerate development timelines for partners.
Nona Biosciences is a wholly owned subsidiary of HBM Holding Limited (HKEX:02142). The HCAb data underlying the model was generated using Nona's proprietary HCAb Harbour Mice platform, described as the world's first clinically validated fully human HCAb transgenic mouse system.