Can human experts predict solubility better than computers?

Abstract

In this study, we design and carry out a survey, asking human experts to predict the aqueous solubility of druglike organic compounds. We investigate whether these experts, drawn largely from the pharmaceutical industry and academia, can match or exceed the predictive power of algorithms. Alongside this, we implement 10 typical machine learning algorithms on the same dataset. The best algorithm, a variety of neural network known as a multi-layer perceptron, gave an RMSE of 0.985 log S units and an R2 of 0.706. We would not have predicted the relative success of this particular algorithm in advance. We found that the best individual human predictor generated an almost identical prediction quality with an RMSE of 0.942 log S units and an R2 of 0.723. The collection of algorithms contained a higher proportion of reasonably good predictors, nine out of ten compared with around half of the humans. We found that, for either humans or algorithms, combining individual predictions into a consensus predictor by taking their median generated excellent predictivity. While our consensus human predictor achieved very slightly better headline figures on various statistical measures, the difference between it and the consensus machine learning predictor was both small and statistically insignificant. We conclude that human experts can predict the aqueous solubility of druglike molecules essentially equally well as machine learning algorithms. We find that, for either humans or algorithms, combining individual predictions into a consensus predictor by taking their median is a powerful way of benefitting from the wisdom of crowds.Electronic supplementary materialThe online version of this article (10.1186/s13321-017-0250-y) contains supplementary material, which is available to authorized users.

Automated De Novo Drug Design: Are We Nearly There Yet?

“…As an example of the still limited predictability of relevant chemical properties,a queous solubility prediction is far from perfect. [76,77] Although there are ongoing efforts to develop improved machine learning methods for this purpose, [78,79] one compound can have several solubilities depending, for example, on multiple crystal forms or amorphous state(s). These important physical and chemical properties must be taken into account for molecular modelling and design, either by explicitly considering them in the model or by providing the suitable training data for implicit knowledge generation.…”

Improved Prediction of Aqueous Solubility of Novel Compounds by Going Deeper With Deep Learning

“…Aqueous solubility is an important physicochemical property of compounds in anti-cancer drug discovery and development, impacting pharmacokinetic properties and formulations (1,2). To facilitate solubility assessment, a number of artificial intelligence (AI) solubility prediction tools have been developed by employing regression and modeling (3,4), machine learning (5)(6)(7)(8)(9), and deep learning (10)(11)(12) methods. These tools have scored impressive performances with high R 2 (e.g., 0.62-0.97) and low RMSE (e.g., 0.29-0.89) values (5,13).…”

Review of the Cosolvency Models for Predicting Drug Solubility in Solvent Mixtures: An Update

“…More experimental and computational efforts are demanded to provide such a tool. It is obvious that the quality of the experimental data is an important factor in providing accurate models (40)(41)(42). To achieve this valuable task, more comprehensive solubility database in mono-and mixed solvents should be generated by the research groups around the world and also more comprehensive and preferably theoretical predictive tools should be provided.…”