Predictive modeling of estrogen receptor agonism, antagonism, and binding activities using machine- and deep-learning approaches

Abstract

As defined by the World Health Organization, an endocrine disruptor is an exogenous substance or mixture that alters function(s) of the endocrine system and consequently causes adverse health effects in an intact organism, its progeny, or (sub)populations. Traditional experimental testing regimens to identify toxicants that induce endocrine disruption can be expensive and time-consuming. Computational modeling has emerged as a promising and cost-effective alternative method for screening and prioritizing potentially endocrine active compounds. The efficient identification of suitable chemical descriptors and machine learning algorithms, including deep learning, is a considerable challenge for computational toxicology studies. Here, we sought to apply classic machine learning algorithms and deep learning approaches to a panel of over 7,500 compounds tested against 18 Toxicity Forecaster (ToxCast) assays related to nuclear estrogen receptor (ERα and ERβ) activity. Three binary fingerprints (Extended Connectivity FingerPrints, Functional Connectivity FingerPrints, and Molecular ACCess System) were used as chemical descriptors in this study. Each descriptor was combined with four machine learning, and two deep learning (normal and multitask neural networks) approaches to construct models for all 18 ER assays. The resulting model performance was evaluated using the area under the receiving operating curve (AUC) values obtained from a five-fold cross-validation procedure. The results showed that individual models have AUC values that range from 0.56 to 0.86. External validation was conducted using two additional sets of compounds (n=592 and n=966) with established interactions with nuclear ER demonstrated through experimentation. An agonist, antagonist, or binding score was determined for each compound by averaging its predicted probabilities in relevant assay models as an external validation, yielding AUC values ranging from 0.63 to 0.91. The results suggest that multitask neural networks offer advantages when modeling mechanistically-related endpoints. Consensus predictions based on the average values of individual models remain the best modeling strategy for computational toxicity evaluations.

Predicting molecular activity on nuclear receptors by multitask neural networks

“…22,23 Multitask neural networks have shown in different QSAR studies to outperform single-task models. 14,22,24,25 The doubtless advantage of the multitask approach, which calculates only one model for several tasks, is that it is cheaper in terms of computational requirements than the traditional single-task QSAR modelling, which implies the calculation of as many models as the tasks. In addition, in multitask modelling, underrepresented tasks can benefit from implicit data augmentation and, thus, gain higher performance.…”

Construction of a Virtual Opioid Bioprofile: A Data-Driven QSAR Modeling Study to Identify New Analgesic Opioids

“…SVM regression models try to find a function of descriptor–activity space that has a limited deviation from the actual activity value for all the training data and at the same time is as flat as possible . These machine learning approaches were tuned to identify the optimal parameters for model performance, as described previously. , …”

“…Individual regression models for each bioassay endpoint were developed using the combination of one type of descriptors (ECFP6, FCFP6, MACCS, rdkit) and one of the modeling approaches ( k NN, RF, SVM), resulting in 12 individual models. The consensus QSAR model, which was generated by averaging predictions of various individual models, was also used in this study. ,, All models were evaluated using a standard 5-fold cross-validation procedure, with 20% of the training set compounds left out for testing purposes during each iteration, as described in previous studies. ,, Each bioassay training set was randomly split into five equal subsets. Four subsets (80% of the total compounds) were used for model training, and the remaining 20% was used to test the resulted model.…”

“…The threshold value to define the AD for a QSAR model using RDKit descriptors was set as the mean plus one-half of the standard deviation of the distribution of distances between each training set compound and its nearest neighbor in the training set, as described in previous studies. , If the Euclidean distance of the test compound from its nearest neighbor in the training set exceeded this threshold, the prediction was considered unreliable. For ECFP6, FCFP6, and MACCS fingerprints, a similarity-based approach was used to define AD, and the threshold was set at a Jaccard similarity of 0.5 . If the Jaccard similarity of the test compound and its nearest neighbor in the training set was less than 0.5, it was considered to be out of AD.…”

How Do Indirect Effects of Contaminants Inform Ecotoxicology? A Review

“…Most research in ecotoxicology has been at the individual-effects level [117], and methods to improve toxicity models continue to be developed. For example, models based on species-level traits and bioenergetics have the potential to improve predictions of species-level responses to various chemicals [114,118], and databases associated with toxic effects that include traditional toxicological information and the results of toxicogenomic or other molecular methods that identify the mode of toxic action are being developed, in some instances with computational modeling [119,120]. Information on direct effects on individuals will improve population models that estimate safe exposure limits for species, representing an important starting place to predict indirect effects.…”