Deep Bayesian Experimental Design for Drug Discovery

AI in Drug Discovery - 1st International Workshop, AIDD 2024, Held in Conjunction with ICANN 2024, Proceedings, pp. 149-159, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) ; Volume 14894 LNCS

Abstract

In drug discovery, prioritizing compounds for testing is an important task. Active learning can assist in this endeavor by prioritizing molecules for label acquisition based on their estimated potential to enhance in-silico models. However, in specialized cases like toxicity modeling, limited dataset sizes can hinder effective training of modern neural networks for representation learning and to perform active learning. In this study, we leverage a transformer-based BERT model pretrained on millions of SMILES to perform active learning. Additionally, we explore different acquisition functions to assess their compatibility with pretrained BERT model. Our results demonstrate that pretrained models enhance active learning outcomes. Furthermore, we observe that active learning selects a higher proportion of positive compounds compared to random acquisition functions, an important advantage, especially in dealing with imbalanced toxicity datasets. Through a comparative analysis, we find that both BALD and EPIG acquisition functions outperform random acquisition, with EPIG exhibiting slightly superior performance over BALD. In summary, our study highlights the effectiveness of active learning in conjunction with pretrained models to tackle the problem of data scarcity.

Description

Keywords

Active learning, Bayesian, BERT, Drug Discovery

DOI

10.1007/978-3-031-72381-0_12

Citation

Masood, M A, Cui, T & Kaski, S 2024, Deep Bayesian Experimental Design for Drug Discovery. in D-A Clevert, M Wand, J Schmidhuber, K Malinovská & I V Tetko (eds), AI in Drug Discovery - 1st International Workshop, AIDD 2024, Held in Conjunction with ICANN 2024, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 14894 LNCS, Springer, pp. 149-159, International Workshop on AI in Drug Discovery, Lugano, Switzerland, 19/09/2024. https://doi.org/10.1007/978-3-031-72381-0_12

Permanent link to this item

https://urn.fi/URN:NBN:fi:aalto-202410236836

Collections

[article-cris] Perustieteiden korkeakoulu / SCI

Show all metadata

Deep Bayesian Experimental Design for Drug Discovery

Access rights

URL

Journal Title

Journal ISSN

Volume Title

Authors

Date

Department

Major/Subject

Mcode

Degree programme

Language

Pages

Series