Bidirectional Learning for Offline Model-based Biological Sequence Design

TL;DR

This paper introduces a novel offline model-based biological sequence design method that leverages pre-trained language models and a bi-level optimization framework to improve sequence scoring and design efficiency.

Contribution

It proposes a new proxy model combining pre-trained LMs with a linear head, and a bi-level optimization framework with adaptive learning rate for better sequence design.

Findings

Effective in DNA and protein sequence design tasks

Outperforms existing methods in sequence optimization accuracy

Demonstrates robustness across different biological datasets

Abstract

Offline model-based optimization aims to maximize a black-box objective function with a static dataset of designs and their scores. In this paper, we focus on biological sequence design to maximize some sequence score. A recent approach employs bidirectional learning, combining a forward mapping for exploitation and a backward mapping for constraint, and it relies on the neural tangent kernel (NTK) of an infinitely wide network to build a proxy model. Though effective, the NTK cannot learn features because of its parametrization, and its use prevents the incorporation of powerful pre-trained Language Models (LMs) that can capture the rich biophysical information in millions of biological sequences. We adopt an alternative proxy model, adding a linear head to a pre-trained LM, and propose a linearization scheme. This yields a closed-form loss and also takes into account the biophysical information in the pre-trained LM. In addition, the forward mapping and the backward mapping play different roles and thus deserve different weights during sequence optimization. To achieve this, we train an auxiliary model and leverage its weak supervision signal via a bi-level optimization framework to effectively learn how to balance the two mappings. Further, by extending the framework, we develop the first learning rate adaptation module \textit{Adaptive}-$\eta$, which is compatible with all gradient-based algorithms for offline model-based optimization. Experimental results on DNA/protein sequence design tasks verify the effectiveness of our algorithm. Our code is available~\href{https://anonymous.4open.science/r/BIB-ICLR2023-Submission/README.md}{here.}