![Sample our Bioscience journals, sign in here to start your access, Latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5925/TOC-Subject-Sample-2021-Bioscience.jpg"})
ABSTRACT
Introduction: Computational modeling has rapidly advanced over the last decades. Recently, machine learning has emerged as a powerful and cost-effective strategy to learn from existing datasets and perform predictions on unseen molecules. Accordingly, the explosive rise of data-driven techniques raises an important question: What confidence can be assigned to molecular property predictions and what techniques can be used?
Areas covered: The authors discuss popular strategies for predicting molecular properties, their corresponding uncertainty sources and methods to quantify uncertainty. First, the authors’ considerations for assessing confidence begin with dataset bias and size, data-driven property prediction and feature design. Next, the authors discuss property simulation via computations of binding affinity in detail. Lastly, they investigate how these uncertainties propagate to generative models, as they are usually coupled with property predictors.
Expert opinion: Computational techniques are paramount to reduce the prohibitive cost of brute-force experimentation during exploration. The authors believe that assessing uncertainty in property prediction models is essential whenever closed-loop drug design campaigns relying on high-throughput virtual screening are deployed. Accordingly, considering sources of uncertainty leads to better-informed validations, more reliable predictions and more realistic expectations of the entire workflow. Overall, this increases confidence in the predictions and, ultimately, accelerates drug design.
Article highlights
For many important properties in drug discovery, only a limited amount of high-quality data is available. Moreover, the data might only be available for specific structural families, which can introduce bias.
Predictive models obtained using inductive inference are never formally correct and inherently uncertain. There are often additional sources of uncertainty that stem from noise or imprecision on target measurements. Many ML methods are available for representing uncertainty.
Property predictors are provided inputs that are considered important for estimating molecular properties.
A key property of interest in drug discovery is the binding affinity of a ligand to a receptor of interest.
Generative models have recently been introduced as techniques for discovering drugs with desired properties. Molecular property predictors are often attached to these generative models.
This box summarizes key points contained in the article.
Abbreviations
ML - machine learning;
ADMET - absorption, distribution, metabolism, excretion and toxicity;
QSAR - quantitative structure-activity relationships;
AD - applicability domain
CP - conformal prediction
GP - gaussian processes
BNN - bayesian neural network
SMILES - simplified molecular input line entry system
Cryo-EM - cryo-electron microscopy
FEP - free energy perturbation
MD - molecular dynamics
VAE - variational autoencoders
GAN - generative adversarial networks
RL - einforcement learning
GAs - genetic algorithms
Acknowledgments
Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the Office of Naval Research.
Declaration of interest
A Aspuru-Guzik is a co-founder and the Chief Visionary Officer at Kebotix Inc. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.