An Active Learning Framework for Data-Efficient, Human-in-the-Loop Enzyme Function Prediction

TL;DR

This paper presents HATTER, an active learning framework that efficiently improves enzyme function prediction by integrating human-in-the-loop annotation, reducing data and computational requirements while maintaining high performance.

Contribution

Introduction of HATTER, a modular active learning framework that combines multiple strategies with human input for scalable, data-efficient enzyme function prediction.

Findings

Active learning matches standard training performance with less data.

Point-based uncertainty sampling methods perform as well or better than complex methods.

Human-in-the-loop active learning accelerates enzyme discovery efficiently.

Abstract

Generalizable protein function prediction is increasingly constrained by the growing mismatch between exponentially expanding sequences of environmental proteins and the comparatively slow accumulation of experimentally verified functional data. Active learning offers a promising path forward for accelerating biological function prediction, by selecting the most informative proteins to experimentally annotate for data-efficient training, yet its potential remains largely unexplored. We introduce HATTER (Human-in-the-loop Adaptive Toolkit for Transferable Enzyme Representations), a modular framework that integrates multiple active learning strategies with human-in-the-loop experimental annotation to efficiently fine tune function prediction models. We compare active learning training to standard supervised training for biological enzyme function prediction, demonstrating that active learning achieves performance comparable to standard training across diverse protein sequence evaluation datasets while requiring fewer model updates, processing less data, and substantially reducing computational cost. Interestingly, point-based uncertainty sampling methods like entropy or margin sampling perform as well or better than more complex acquisition functions such as bayesian sampling or BALD, highlighting the relative importance of sequence diversity in training datasets and model architecture design. These results demonstrate that human-in-the-loop active learning can efficiently accelerate enzyme discovery, providing a flexible platform for adaptive, scalable, and expert-guided protein function prediction.