A prediction model of drug-induced ototoxicity developed by an optimal support vector machine (SVM) method

Ototoxicity: a high risk to auditory function that needs to be monitored in drug development

Hearing loss constitutes a major global health concern impacting approximately 1.5 billion people worldwide. Its incidence is undergoing a substantial surge with some projecting that by 2050, a quarter of the global population will experience varying degrees of hearing deficiency. Environmental factors such as aging, exposure to loud noise, and the intake of ototoxic medications are implicated in the onset of acquired hearing loss. Ototoxicity resulting in inner ear damage is a leading cause of acquired hearing loss worldwide. This could be minimized or avoided by early testing of hearing functions in the preclinical phase of drug development. While the assessment of ototoxicity is well defined for drug candidates in the hearing field - required for drugs that are administered by the otic route and expected to reach the middle or inner ear during clinical use - ototoxicity testing is not required for all other therapeutic areas. Unfortunately, this has resulted in more than 200 ototoxic marketed medications. The aim of this publication is to raise awareness of drug-induced ototoxicity and to formulate some recommendations based on available guidelines and own experience. Ototoxicity testing programs should be adapted to the type of therapy, its indication (targeting the ear or part of other medications classes being potentially ototoxic), and the number of assets to test. For multiple molecules and/or multiple doses, screening options are available: in vitro (otic cell assays), ex vivo (cochlear explant), and in vivo (in zebrafish). In assessing the ototoxicity of a candidate drug, it is good practice to compare its ototoxicity to that of a well-known control drug of a similar class. Screening assays provide a streamlined and rapid method to know whether a drug is generally safe for inner ear structures. Mammalian animal models provide a more detailed characterization of drug ototoxicity, with a possibility to localize and quantify the damage using functional, behavioral, and morphological read-outs. Complementary histological measures are routinely conducted notably to quantify hair cells loss with cochleogram. Ototoxicity studies can be performed in rodents (mice, rats), guinea pigs and large species. However, in undertaking, or at the very least attempting, all preclinical investigations within the same species, is crucial. This encompasses starting with pharmacokinetics and pharmacology efficacy studies and extending through to toxicity studies. In life read-outs include Auditory Brainstem Response (ABR) and Distortion Product OtoAcoustic Emissions (DPOAE) measurements that assess the activity and integrity of sensory cells and the auditory nerve, reflecting sensorineural hearing loss. Accurate, reproducible, and high throughput ABR measures are fundamental to the quality and success of these preclinical trials. As in humans, in vivo otoscopic evaluations are routinely carried out to observe the tympanic membrane and auditory canal. This is often done to detect signs of inflammation. The cochlea is a tonotopic structure. Hair cell responsiveness is position and frequency dependent, with hair cells located close to the cochlea apex transducing low frequencies and those at the base transducing high frequencies. The cochleogram aims to quantify hair cells all along the cochlea and consequently determine hair cell loss related to specific frequencies. This measure is then correlated with the ABR & DPOAE results. Ototoxicity assessments evaluate the impact of drug candidates on the auditory and vestibular systems, de-risk hearing loss and balance disorders, define a safe dose, and optimize therapeutic benefits. These types of studies can be initiated during early development of a therapeutic solution, with ABR and otoscopic evaluations. Depending on the mechanism of action of the compound, studies can include DPOAE and cochleogram. Later in the development, a GLP (Good Laboratory Practice) ototoxicity study may be required based on otic related route of administration, target, or known potential otic toxicity.

Machine Learning Methods for Small Data Challenges in Molecular Science

Small data are often used in scientific and engineering research due to the presence of various constraints, such as time, cost, ethics, privacy, security, and technical limitations in data acquisition. However, big data have been the focus for the past decade, small data and their challenges have received little attention, even though they are technically more severe in machine learning (ML) and deep learning (DL) studies. Overall, the small data challenge is often compounded by issues, such as data diversity, imputation, noise, imbalance, and high-dimensionality. Fortunately, the current big data era is characterized by technological breakthroughs in ML, DL, and artificial intelligence (AI), which enable data-driven scientific discovery, and many advanced ML and DL technologies developed for big data have inadvertently provided solutions for small data problems. As a result, significant progress has been made in ML and DL for small data challenges in the past decade. In this review, we summarize and analyze several emerging potential solutions to small data challenges in molecular science, including chemical and biological sciences. We review both basic machine learning algorithms, such as linear regression, logistic regression (LR), k-nearest neighbor (KNN), support vector machine (SVM), kernel learning (KL), random forest (RF), and gradient boosting trees (GBT), and more advanced techniques, including artificial neural network (ANN), convolutional neural network (CNN), U-Net, graph neural network (GNN), Generative Adversarial Network (GAN), long short-term memory (LSTM), autoencoder, transformer, transfer learning, active learning, graph-based semi-supervised learning, combining deep learning with traditional machine learning, and physical model-based data augmentation. We also briefly discuss the latest advances in these methods. Finally, we conclude the survey with a discussion of promising trends in small data challenges in molecular science.

Exploring different computational approaches for effective diagnosis of breast cancer

Breast cancer has been identified as one among the top causes of female death worldwide. According to recent research, earlier detection plays an important role toward fortunate medicaments and thus, decreasing the mortality rate due to breast cancer among females. This review provides a fleeting summary involving traditional diagnostic procedures from the past and today, and also modern computational tools that have greatly aided in the identification of breast cancer. Computational techniques involving different algorithms such as Support vector machines, deep learning techniques and robotics are popular among the academicians for detection of breast cancer. They discovered that Convolutional neural network was a common option for categorization among such approaches. Deep learning techniques are evaluated using performance indicators such as accuracy, sensitivity, specificity, or measure. Furthermore, molecular docking, homology modeling and Molecular dynamics Simulation gives a road map for future discussions about developing improved early detection approaches that holds greater potential in increasing the survival rate of cancer patients. The different computational techniques can be a new dominion among researchers and combating the challenges associated with breast cancer.

Hair cell toxicology: With the help of a little fish

Hearing or balance loss are disabling conditions that have a serious impact in those suffering them, especially when they appear in children. Their ultimate cause is frequently the loss of function of mechanosensory hair cells in the inner ear. Hair cells can be damaged by environmental insults, like noise or chemical agents, known as ototoxins. Two of the most common ototoxins are life-saving medications: cisplatin against solid tumors, and aminoglycoside antibiotics to treat infections. However, due to their localization inside the temporal bone, hair cells are difficult to study in mammals. As an alternative animal model, zebrafish larvae have hair cells similar to those in mammals, some of which are located in a fish specific organ on the surface of the skin, the lateral line. This makes them easy to observe in vivo and readily accessible for ototoxins or otoprotective substances. These features have made possible advances in the study of the mechanisms mediating ototoxicity or identifying new potential ototoxins. Most importantly, the small size of the zebrafish larvae has allowed screening thousands of molecules searching for otoprotective agents in a scale that would be highly impractical in rodent models. The positive hits found can then start the long road to reach clinical settings to prevent hearing or balance loss.

Mechanisms of Aminoglycoside- and Cisplatin-Induced Ototoxicity

Purpose This review article summarizes our current understanding of the mechanisms underlying acquired hearing loss from hospital-prescribed medications that affects as many as 1 million people each year in Western Europe and North America. Yet, there are currently no federally approved drugs to prevent or treat the debilitating and permanent hearing loss caused by the life-saving platinum-based anticancer drugs or the bactericidal aminoglycoside antibiotics. Hearing loss has long-term impacts on quality-of-life measures, especially in young children and older adults. This review article also highlights some of the current knowledge gaps regarding iatrogenic causes of hearing loss. Conclusion Further research is urgently needed to further refine clinical practice and better ameliorate iatrogenic drug-induced hearing loss.

Machine learning - Predicting Ames mutagenicity of small molecules

In modern drug discovery, detection of a compound's potential mutagenicity is crucial. However, the traditional method of mutagenicity detection using the Ames test is costly and time consuming as the compounds need to be synthesised and then tested and the results are not always accurate and reproducible. Therefore, it would be advantageous to develop robust in silico models which can accurately predict the mutagenicity of a compound prior to synthesis to overcome the inadequacies of the Ames test. After curation of a previously defined compound mutagenicity library, over 5000 molecules had their chemical fingerprints and molecular properties calculated. Using 8 classification modelling algorithms, including support vector machine (SVM), random forest (RF) and extreme gradient boosting (XGB), a total of 112 predictive models have been constructed. Their performance has been assessed using 10-fold cross validation and a hold-out test set and some of the top performing models have been assessed using the y-randomisation approach. As a result, we have found SVM and XGB models to have good performance during the 10-fold cross validation (AUROC >0.90, sensitivity >0.85, specificity >0.75, balanced accuracy >0.80, Kappa >0.65) and on the test set (AUROC >0.65, sensitivity >0.65, specificity >0.60, balanced accuracy >0.65, Kappa >0.30). We have also identified molecular properties that are the most influential for mutagenicity prediction when combined with chemical molecular fingerprints. Using the Class A mutagenic compounds from the Ames/QSAR International Challenge Project, we were able to verify our models perform better, predicting more mutagens correctly then the StarDrop Ames mutagenicity prediction and TEST mutagenicity prediction.

Detecting Novel Ototoxins and Potentiation of Ototoxicity by Disease Settings

Over 100 drugs and chemicals are associated with permanent hearing loss, tinnitus, and vestibular deficits, collectively known as ototoxicity. The ototoxic potential of drugs is rarely assessed in pre-clinical drug development or during clinical trials, so this debilitating side-effect is often discovered as patients begin to report hearing loss. Furthermore, drug-induced ototoxicity in adults, and particularly in elderly patients, may go unrecognized due to hearing loss from a variety of etiologies because of a lack of baseline assessments immediately prior to novel therapeutic treatment. During the current pandemic, there is an intense effort to identify new drugs or repurpose FDA-approved drugs to treat COVID-19. Several potential COVID-19 therapeutics are known ototoxins, including chloroquine (CQ) and lopinavir-ritonavir, demonstrating the necessity to identify ototoxic potential in existing and novel medicines. Furthermore, several factors are emerging as potentiators of ototoxicity, such as inflammation (a hallmark of COVID-19), genetic polymorphisms, and ototoxic synergy with co-therapeutics, increasing the necessity to evaluate a drug's potential to induce ototoxicity under varying conditions. Here, we review the potential of COVID-19 therapies to induce ototoxicity and factors that may compound their ototoxic effects. We then discuss two models for rapidly detecting the potential for ototoxicity: mammalian auditory cell lines and the larval zebrafish lateral line. These models offer considerable value for pre-clinical drug development, including development of COVID-19 therapies. Finally, we show the validity of in silico screening for ototoxic potential using a computational model that compares structural similarity of compounds of interest with a database of known ototoxins and non-ototoxins. Preclinical screening at in silico, in vitro, and in vivo levels can provide an earlier indication of the potential for ototoxicity and identify the subset of candidate therapeutics for treating COVID-19 that need to be monitored for ototoxicity as for other widely-used clinical therapeutics, like aminoglycosides and cisplatin.

In silico prediction of drug-induced ototoxicity using machine learning and deep learning methods

Drug-induced ototoxicity has become a serious global problem, because of leading to deafness in hundreds of thousands of people every year. It always results from exposure to drugs or environmental chemicals that cause the impairment and degeneration of the inner ear. Herein, we focused on the in silico modeling of drug-induced ototoxicity of chemicals. We collected 1,102 ototoxic medications and 1,705 non-ototoxic drugs. Based on the data set, a series of computational models were developed with different traditional machine learning and deep learning algorithms implemented on an online chemical database and modeling environment. Six ML models performed best on 5-fold cross-validation and test set. A consensus model was developed with the best individual models. These models were further validated with an external validation. The consensus model showed best predictive ability, with high accuracy of 0.95 on test set and 0.90 on validation set. The consensus model and the data sets used for model development are available at https://ochem.eu/model/46566321. Besides, 16 structural alerts responsible for drug-induced ototoxicity were identified. We hope the results could provide meaningful knowledge and useful tools for ototoxicity evaluation in drug discovery and environmental risk assessment.

Machine Learning in Predictive Toxicology: Recent Applications and Future Directions for Classification Models

In recent times, machine learning has become increasingly prominent in predictive toxicology as it has shifted from in vivo studies toward in silico studies. Currently, in vitro methods together with other computational methods such as quantitative structure-activity relationship modeling and absorption, distribution, metabolism, and excretion calculations are being used. An overview of machine learning and its applications in predictive toxicology is presented here, including support vector machines (SVMs), random forest (RF) and decision trees (DTs), neural networks, regression models, naïve Bayes, k-nearest neighbors, and ensemble learning. The recent successes of these machine learning methods in predictive toxicology are summarized, and a comparison of some models used in predictive toxicology is presented. In predictive toxicology, SVMs, RF, and DTs are the dominant machine learning methods due to the characteristics of the data available. Lastly, this review describes the current challenges facing the use of machine learning in predictive toxicology and offers insights into the possible areas of improvement in the field.

A Hybrid Landslide Displacement Prediction Method Based on CEEMD and DTW-ACO-SVR-Cases Studied in the Three Gorges Reservoir Area

Accurately predicting the surface displacement of the landslide is important and necessary. However, most of the existing research has ignored the frequency component of inducing factors and how it affects the landslide deformation. Therefore, a hybrid displacement prediction model based on time series theory and various intelligent algorithms was proposed in this paper to study the effect of frequency components. Firstly, the monitoring displacement of landslide from the Three Gorges Reservoir area (TGRA) was decomposed into the trend and periodic components by complete ensemble empirical mode decomposition (CEEMD). The trend component can be predicted by the least square method. Then, time series of inducing factors like rainfall and reservoir level was reconstructed into high frequency components and low frequency components with CEEMD and t-test, respectively. The dominant factors were selected by the method of dynamic time warping (DTW) from the frequency components and other common factors (e.g., current monthly rainfall). Finally, the ant colony optimization-based support vector machine regression (ACO-SVR) is utilized for prediction purposes in the TGRA. The results demonstrate that after considering the frequency components of landslide-induced factors, the accuracy of the displacement prediction model based on ACO-SVR is better than that of other models based on SVR and GA-SVR.

Performance of machine-learning algorithms to pattern recognition and classification of hearing impairment in Brazilian farmers exposed to pesticide and/or cigarette smoke

The use of pesticides has been increasing in agriculture, leading to a public health problem. The aim of this study was to evaluate ototoxic effects in farmers who were exposed to cigarette smoke and/or pesticides and to identify possible classification patterns in the exposure groups. The sample included 127 participants of both sexes aged between 18 and 39, who were divided into the following four groups: control group (CG), smoking group (SG), pesticide group (PG), and smoking + pesticide group (SPG). Meatoscopy, pure tone audiometry, logoaudiometry, high-frequency thresholds, and immittance testing were performed. Data were evaluated by artificial neural network (ANN), K-nearest neighbors (K-NN), and support vector machine (SVM). There was symmetry between the right and left ears, an increase in the incidence of hearing loss at high frequency and of downward sloping audiometric curve configuration, and alteration of stapedial reflex in the three exposed groups. The machine-learning classifiers achieved good classification performance (control and exposed). The best classification results occur in high type (I and II) datasets (about 90% accuracy) in k-NN test. It is concluded that both xenobiotic substances have ototoxic potential; however, their combined use does not present additive or potentiating effects recognizable by the algorithms.

ADMET Evaluation in Drug Discovery. Part 17: Development of Quantitative and Qualitative Prediction Models for Chemical-Induced Respiratory Toxicity

As a dangerous end point, respiratory toxicity can cause serious adverse health effects and even death. Meanwhile, it is a common and traditional issue in occupational and environmental protection. Pharmaceutical and chemical industries have a strong urge to develop precise and convenient computational tools to evaluate the respiratory toxicity of compounds as early as possible. Most of the reported theoretical models were developed based on the respiratory toxicity data sets with one single symptom, such as respiratory sensitization, and therefore these models may not afford reliable predictions for toxic compounds with other respiratory symptoms, such as pneumonia or rhinitis. Here, based on a diverse data set of mouse intraperitoneal respiratory toxicity characterized by multiple symptoms, a number of quantitative and qualitative predictions models with high reliability were developed by machine learning approaches. First, a four-tier dimension reduction strategy was employed to find an optimal set of 20 molecular descriptors for model building. Then, six machine learning approaches were used to develop the prediction models, including relevance vector machine (RVM), support vector machine (SVM), regularized random forest (RRF), extreme gradient boosting (XGBoost), naïve Bayes (NB), and linear discriminant analysis (LDA). Among all of the models, the SVM regression model shows the most accurate quantitative predictions for the test set (q²_ext = 0.707), and the XGBoost classification model achieves the most accurate qualitative predictions for the test set (MCC of 0.644, AUC of 0.893, and global accuracy of 82.62%). The application domains were analyzed, and all of the tested compounds fall within the application domain coverage. We also examined the structural features of the compounds and important fragments with large prediction errors. In conclusion, the SVM regression model and the XGBoost classification model can be employed as accurate prediction tools for respiratory toxicity.

ADMET evaluation in drug discovery: 15. Accurate prediction of rat oral acute toxicity using relevance vector machine and consensus modeling

Background

Determination of acute toxicity, expressed as median lethal dose (LD₅₀), is one of the most important steps in drug discovery pipeline. Because in vivo assays for oral acute toxicity in mammals are time-consuming and costly, there is thus an urgent need to develop in silico prediction models of oral acute toxicity.

Results

In this study, based on a comprehensive data set containing 7314 diverse chemicals with rat oral LD₅₀ values, relevance vector machine (RVM) technique was employed to build the regression models for the prediction of oral acute toxicity in rate, which were compared with those built using other six machine learning approaches, including k-nearest-neighbor regression, random forest (RF), support vector machine, local approximate Gaussian process, multilayer perceptron ensemble, and eXtreme gradient boosting. A subset of the original molecular descriptors and structural fingerprints (PubChem or SubFP) was chosen by the Chi squared statistics. The prediction capabilities of individual QSAR models, measured by q_ext² for the test set containing 2376 molecules, ranged from 0.572 to 0.659.

Conclusion

Considering the overall prediction accuracy for the test set, RVM with Laplacian kernel and RF were recommended to build in silico models with better predictivity for rat oral acute toxicity. By combining the predictions from individual models, four consensus models were developed, yielding better prediction capabilities for the test set (q_ext² = 0.669–0.689). Finally, some essential descriptors and substructures relevant to oral acute toxicity were identified and analyzed, and they may be served as property or substructure alerts to avoid toxicity. We believe that the best consensus model with high prediction accuracy can be used as a reliable virtual screening tool to filter out compounds with high rat oral acute toxicity. Graphical abstract

Predicting selective liver X receptor β agonists using multiple machine learning methods

The classification models for predicting selective LXRβ agonists were firstly established using multiple machine learning methods. The top models can predict selective LXRβ agonists with chemical structure diversity.