Machine learning applications in drug development

Abstract

Due to the huge amount of biological and medical data available today, along with well-established machine learning algorithms, the design of largely automated drug development pipelines can now be envisioned. These pipelines may guide, or speed up, drug discovery; provide a better understanding of diseases and associated biological phenomena; help planning preclinical wet-lab experiments, and even future clinical trials. This automation of the drug development process might be key to the current issue of low productivity rate that pharmaceutical companies currently face. In this survey, we will particularly focus on two classes of methods: sequential learning and recommender systems, which are active biomedical fields of research.

Keywords: Adaptive clinical trial; Bayesian optimization; Collaborative filtering; Drug discovery; Drug repurposing; Multi-armed bandit.

Graphical abstract

Fig. 1

Representation of the four stages of drug development, along with Phase IV, which occurs after the start of drug marketing.

Fig. 2

Evolution of average development cost in a cohort of 12 major phamaceutical labs, in millions of dollars, between 2010 and 2018 .

Fig. 3

A K-armed bandit, where the learning agent interacts with its environment by sequentially selecting arms, and updating its strategy using the observations it obtains.

Fig. 4

Generative Adversarial Networks for Drug Discovery. A Generative Adversarial Network is a set of two neural networks, the Generator and the Discriminator. These two networks are trained at the same time.

Fig. 5

Randomized Clinical Trial (RCT) versus Adaptive Clinical Trial (ACT) for Phase III. A Randomized Clinical Trial (RCT) “randomly” assign patients to treatment arms (ensuring balance of covariates of interest) before testing, whereas an Adapted Clinical Trial sequentially assigns patients to treatment arms according to previous testing results.