Predicting Dose-Range Chemical Toxicity using Novel Hybrid Deep Machine-Learning Method

Abstract

Humans are exposed to thousands of chemicals, including environmental chemicals. Unfortunately, little is known about their potential toxicity, as determining the toxicity remains challenging due to the substantial resources required to assess a chemical in vivo. Here, we present a novel hybrid neural network (HNN) deep learning method, called HNN-Tox, to predict chemical toxicity at different doses. To develop a hybrid HNN-Tox method, we combined two neural network frameworks, the Convolutional Neural Network (CNN) and the multilayer perceptron (MLP)-type feed-forward neural network (FFNN). Combining the CNN and FCNN in the field of environmental chemical toxicity prediction is a novel approach. We developed several binary and multiclass classification models to assess dose-range chemical toxicity that is trained based on thousands of chemicals with known toxicity. The performance of the HNN-Tox was compared with other machine-learning methods, including Random Forest (RF), Bootstrap Aggregation (Bagging), and Adaptive Boosting (AdaBoost). We also analyzed the model performance dependency on varying features, descriptors, dataset size, route of exposure, and toxic dose. The HNN-Tox model, trained on 59,373 chemicals annotated with known LD50 and routes of exposure, maintained its predictive ability with an accuracy of 84.9% and 84.1%, even after reducing the descriptor size from 318 to 51, and the area under the ROC curve (AUC) was 0.89 and 0.88, respectively. Further, we validated the HNN-Tox with several external toxic chemical datasets on a large scale. The HNN-Tox performed optimally or better than the other machine-learning methods for diverse chemicals. This study is the first to report a large-scale prediction of dose-range chemical toxicity with varying features. The HNN-Tox has broad applicability in predicting toxicity for diverse chemicals and could serve as an alternative methodology approach to animal-based toxicity assessment.

Predicting Dose-Dependent Carcinogenicity of Chemical Mixtures Using a Novel Hybrid Neural Network Framework and Mathematical Approach

“…A vector was created for the SMILES of each compound by converting each character in the SMILES string to its corresponding index in the dictionary. The resulting vector for the SMILES was padded with zeros or truncated so that it was of uniform length, L. We previously described in detail the SMILES preprocessing of chemicals [ 76 , 77 , 78 ].…”

Quantum-to-Classical Neural Network Transfer Learning Applied to Drug Toxicity Prediction

“…CNNs have been found to be successful for use in toxicity prediction tasks in the literature − and are very popular QML analogues of classical architectures. − The popularity of QCNNs partly stems from only needing to operate on small sections of input data at a time, rendering them feasible for NISQ devices. We employ a quantum-classical architecture where the first convolutional layer is replaced by a quantum circuit and is subsequently followed by classical pooling, a classical convolutional layer, and a fully connected layer, as depicted in Figure .…”

Gap-Δenergy, a New Metric of the Bond Energy State, Assisting to Predict Molecular Toxicity

“…The high attrition rate of small-molecule drug candidates is primarily attributed to safety and toxicology in the preclinical or clinical phase . Numerous in vitro and in vivo experiments are required to assess drug toxicity, which are time-consuming and expensive, and in vivo animal tests may arouse ethical concerns . Furthermore, the differences in physiology and genetics between humans and animal models can cause the inapplicability of toxicity prediction to humans. , Studies have shown that the absence of toxicity observations from preclinical animal tests does not imply harmlessness for humans. , Therefore, it is crucial to utilize computational toxicology to assist in high-throughput drug toxicity prediction.…”

“… 5 Numerous in vitro and in vivo experiments are required to assess drug toxicity, which are time-consuming and expensive, and in vivo animal tests may arouse ethical concerns. 6 Furthermore, the differences in physiology and genetics between humans and animal models can cause the inapplicability of toxicity prediction to humans. 7 , 8 Studies have shown that the absence of toxicity observations from preclinical animal tests does not imply harmlessness for humans.…”