Investigating Active Learning and Meta-Learning for Iterative Peptide Design

TL;DR

This paper evaluates active learning and meta-learning techniques to optimize peptide experimental design, finding meta-learning improves accuracy but active learning does not outperform random selection, with mixed results when combined.

Contribution

The study introduces a peptide benchmark database and assesses the effectiveness of active learning and meta-learning methods in reducing experimental efforts.

Findings

Meta-learning with Reptile improves average accuracy.

Active learning methods tested did not outperform random selection.

Combining meta-learning with active learning yields inconsistent benefits.

Abstract

Often the development of novel functional peptides is not amenable to high throughput or purely computational screening methods. Peptides must be synthesized one at a time in a process that does not generate large amounts of data. One way this method can be improved is by ensuring that each experiment provides the best improvement in both peptide properties and predictive modeling accuracy. Here, we study the effectiveness of active learning, optimizing experiment order, and meta-learning, transferring knowledge between contexts, to reduce the number of experiments necessary to build a predictive model. We present a multi-task benchmark database of peptides designed to advance these methods for experimental design. Each task is binary classification of peptides represented as a sequence string. We find neither active learning method tested to be better than random choice. The meta-learning method Reptile was found to improve average accuracy across datasets. Combining meta-learning with active learning offers inconsistent benefits.