Search papers, labs, and topics across Lattice.
This paper introduces ABOPD, an antibody design framework that utilizes on-policy distillation to enhance the generation of complementarity-determining regions (CDRs) in antibodies. By incorporating privileged native geometry during training, ABOPD effectively supervises the states encountered in the model's denoising trajectories, addressing the limitations of standard denoising training methods. The key result shows a significant improvement in structural recovery for CDR-H3 generation, achieving a reduction in RMSD from 2.37 脜 to 1.95 脜, thus paving the way for more accurate protein design.
ABOPD achieves a remarkable 0.42 脜 reduction in RMSD for antibody CDR design, setting a new standard for structural fidelity in protein generation.
Antibodies are essential therapeutic molecules, and their complementarity-determining regions (CDRs) form the primary antigen-recognition interface. Recent protein generative models have demonstrated broad capabilities in biomolecular design, yet post-training strategies for downstream objectives remain limited. Standard denoising training operates on noisy states obtained by perturbing native structures, whereas recursive generation proceeds through model-generated intermediate states. For flexible antibody CDR loops such as CDR-H3, this mismatch can allow backbone deviations to accumulate along the denoising trajectory and compromise antigen-facing loop geometry. We introduce ABOPD, an antibody design framework based on on-policy distillation that leverages privileged native geometry during training to supervise states visited along the model's own denoising trajectories. With this fine-grained structural supervision, ABOPD substantially improves structural recovery on RAbD CDR-H3 generation, reducing RMSD by 0.42 {\AA} (from 2.37 {\AA} to 1.95 {\AA}) and outperforming supervised fine-tuning and offline distillation controls, offering a path to higher-fidelity protein design.
