PropertyDAG: Multi-objective Bayesian optimization of partially ordered, mixed-variable properties for biological sequence design

TL;DR

PropertyDAG enhances Bayesian optimization for biological sequence design by incorporating hierarchical, partial order dependencies among objectives, enabling more realistic and prioritized multi-objective optimization in complex biological tasks.

Contribution

It introduces a novel framework that integrates partial ordering of objectives into Bayesian optimization, addressing hierarchical dependencies often present in biological sequence design.

Findings

Improves optimization performance in simulated active learning tasks.

Effectively models hierarchical dependencies among objectives.

Demonstrates applicability to real-world antibody design.

Abstract

Bayesian optimization offers a sample-efficient framework for navigating the exploration-exploitation trade-off in the vast design space of biological sequences. Whereas it is possible to optimize the various properties of interest jointly using a multi-objective acquisition function, such as the expected hypervolume improvement (EHVI), this approach does not account for objectives with a hierarchical dependency structure. We consider a common use case where some regions of the Pareto frontier are prioritized over others according to a specified $\textit{partial ordering}$ in the objectives. For instance, when designing antibodies, we would like to maximize the binding affinity to a target antigen only if it can be expressed in live cell culture -- modeling the experimental dependency in which affinity can only be measured for antibodies that can be expressed and thus produced in viable quantities. In general, we may want to confer a partial ordering to the properties such that each property is optimized conditioned on its parent properties satisfying some feasibility condition. To this end, we present PropertyDAG, a framework that operates on top of the traditional multi-objective BO to impose this desired ordering on the objectives, e.g. expression $\rightarrow$ affinity. We demonstrate its performance over multiple simulated active learning iterations on a penicillin production task, toy numerical problem, and a real-world antibody design task.