Pareto-Optimal Energy Alignment for Designing Nature-Like Antibodies

We present a three-stage framework for training deep learning models specializing in antibody sequence-structure co-design. We first pre-train a language model using millions of antibody sequence data. Then, we employ the learned representations to guide the training of a diffusion model for joint optimization over both sequence and structure of antibodies. During the final alignment stage, we optimize the model to favor antibodies with low repulsion and high attraction to the antigen binding site, enhancing the rationality and functionality of the designs. To mitigate conflicting energy preferences, we extend AbDPO (Antibody Direct Preference Optimization) to guide the model toward Pareto optimality under multiple energybased alignment objectives. Furthermore, we adopt an iterative learning paradigm with temperature scaling, enabling the model to benefit from diverse online datasets without requiring additional data. In practice, our proposed methods achieve high stability and efficiency in producing a better Pareto front of antibody designs compared to top samples generated by baselines and previous alignment techniques.

Through extensive experiments, we showcase the superior performance of our methods in generating nature-like antibodies with high binding affinity.

In summary, our main contributions are as follows:

• We devise the first three-stage training framework for antibody sequence-structure co-design, consisting of pre-training, transferring, and alignment.

• We propose an efficient multi-objective alignment algorithm with online exploration which consistently produces a better Pareto front of models in terms of energy without extra data.

• Our approach achieves state-of-the-art performance in generating more natural-like antibodies with better rationality and functionality.