Avoiding hERG-liability in drug design via synergetic combinations of different (Q)SAR methodologies and data sources: a case study in an industrial setting

Thierry Hanser,Mireille Krier,Fabian P Steinmetz,Friedrich Rippmann,Jeffrey Plante

doi:10.1186/s13321-019-0334-y

Thierry Hanser, Mireille Krier + Show 3 more

Open Access

https://doi.org/10.1186/s13321-019-0334-y

Copy DOI

Abstract

In this paper, we explore the impact of combining different in silico prediction approaches and data sources on the predictive performance of the resulting system. We use inhibition of the hERG ion channel target as the endpoint for this study as it constitutes a key safety concern in drug development and a potential cause of attrition. We will show that combining data sources can improve the relevance of the training set in regard of the target chemical space, leading to improved performance. Similarly we will demonstrate that combining multiple statistical models together, and with expert systems, can lead to positive synergistic effects when taking into account the confidence in the predictions of the merged systems. The best combinations analyzed display a good hERG predictivity. Finally, this work demonstrates the suitability of the SOHN methodology for building models in the context of receptor based endpoints like hERG inhibition when using the appropriate pharmacophoric descriptors.

Highlights

The inhibition of the human ether-a-go-go ion channel may cause QT interval prolongation, which eventually can result in torsades de pointes (TdP) [1] and even death
It is highly desirable to have a good means of predicting human ether-a-go-go (hERG) activity, and for this purpose in silico systems provide a low cost solution that can be applied to the large datasets in early drug discovery
We examine the combination of different models, including both a traditional QSAR random forest (RF) model and an expert rule-based system along with the newly introduced self-organising hypothesis networks (SOHN) model

Summary

Introduction

The inhibition of the human ether-a-go-go (hERG) ion channel may cause QT interval prolongation, which eventually can result in torsades de pointes (TdP) [1] and even death. Cardiotoxicity caused by the inhibition of hERG is a major liability within the drug development process. To avoid such a severe adverse effect, it makes good sense to screen all potential drug candidates for risk against blocking the hERG channel. It is highly desirable to have a good means of predicting hERG activity, and for this purpose in silico systems provide a low cost solution that can be applied to the large datasets in early drug discovery. A large number of different models have been developed. We present a new model, and a thorough comparison of its results using both public training data as well as mixing it with

Objectives

Methods

Results