Bayesian neural network with pretrained protein embedding enhances prediction accuracy of drug-protein interaction

TL;DR

This paper introduces a novel deep learning framework combining pretrained protein embeddings and Bayesian neural networks to improve drug-protein interaction prediction accuracy, especially with limited labeled data, and utilizes uncertainty quantification for better screening.

Contribution

The paper presents a new approach integrating transfer learning with pretrained protein embeddings and Bayesian neural networks for enhanced DPI prediction with small datasets.

Findings

Model outperforms previous baselines in DPI prediction.

Quantified uncertainty correlates with prediction confidence.

Uncertainty estimation aids in screening DPI data points.

Abstract

The characterization of drug-protein interactions is crucial in the high-throughput screening for drug discovery. The deep learning-based approaches have attracted attention because they can predict drug-protein interactions without trial-and-error by humans. However, because data labeling requires significant resources, the available protein data size is relatively small, which consequently decreases model performance. Here we propose two methods to construct a deep learning framework that exhibits superior performance with a small labeled dataset. At first, we use transfer learning in encoding protein sequences with a pretrained model, which trains general sequence representations in an unsupervised manner. Second, we use a Bayesian neural network to make a robust model by estimating the data uncertainty. As a result, our model performs better than the previous baselines for predicting drug-protein interactions. We also show that the quantified uncertainty from the Bayesian inference is related to the confidence and can be used for screening DPI data points.