Predictive Multitask Deep Neural Network Models for ADME-Tox Properties: Learning from Large Data Sets

Abstract

Successful drug discovery projects require control and optimization of compound properties related to pharmacokinetics, pharmacodynamics, and safety. While volume and chemotype coverage of public and corporate ADME-Tox (absorption, distribution, excretion, metabolism, and toxicity) databases are constantly growing, deep neural nets (DNN) emerged as transformative artificial intelligence technology to analyze those challenging data. Relevant features are automatically identified, while appropriate data can also be combined to multitask networks to evaluate hidden trends among multiple ADME-Tox parameters for implicitly correlated data sets. Here we describe a novel, fully industrialized approach to parametrize and optimize the setup, training, application, and visual interpretation of DNNs to model ADME-Tox data. Investigated properties include microsomal lability in different species, passive permeability in Caco-2/TC7 cells, and logD. Statistical models are developed using up to 50 000 compounds from public or corporate databases. Both the choice of DNN hyperparameters and the type and quantity of molecular descriptors were found to be important for successful DNN modeling. Alternate learning of multiple ADME-Tox properties, resulting in a multitask approach, performs statistically superior on most studied data sets in comparison to DNN single-task models and also provides a scalable method to predict ADME-Tox properties from heterogeneous data. For example, predictive quality using external validation sets was improved from R 2 of 0.6 to 0.7 comparing single-task and multitask DNN networks from human metabolic lability data. Besides statistical evaluation, a new visualization approach is introduced to interpret DNN models termed “response map”, which is useful to detect local property gradients based on structure fragmentation and derivatization. This method is successfully applied to visualize fragmental contributions to guide further design in drug discovery programs, as illustrated by CRCX3 antagonists and renin inhibitors, respectively.

Are 2D fingerprints still valuable for drug discovery?

“…The multitask (MT) learning technique has achieved much success in qualitative Merck and Tox21 prediction challenges. [45][46][47][48] In the MT framework, multiple tasks share the same hidden layers. However, the output layer is attached to different tasks.…”

In silico toxicology: From structure–activity relationships towards deep learning and adverse outcome pathways

“…Many studies already highlighted that multitask networks exhibit better predictivity than single task networks. The application so far was done for modeling of acute toxicity, reactivity, and ADME‐Tox properties . All this work highlights, that without the need to use descriptors or fingerprints to transform a molecule into a suitable input of machine learning models, in silico toxicology can benefit greatly.…”

“…However, interesting conclusions were drawn by Mayr et al, Gini et al, and Xu and coworkers who could show that networks can learn representations which are comparable to structural alerts . Wenzel and coworkers introduced response maps to highlight important features used by the network …”

Transforming cancer drug discovery with Big Data and AI

“…More recently, there have been promising applications of AI methods to ADMET prediction. For example, the recent use of transfer and multitask learning in predicting pharmacokinetic parameters [38] or the application of DL to ADMET prediction [39]. Moreover, public-private partnerships such as the eTOX consortium have been created to recover legacy toxicological data from big pharmaceutical companies and develop better predictive models [40].…”