A compound attributes-based predictive model for drug induced liver injury in humans

Abstract

Drug induced liver injury (DILI) is one of the key safety concerns in drug development. To assess the likelihood of drug candidates with potential adverse reactions of liver, we propose a compound attributes-based approach to predicting hepatobiliary disorders that are routinely reported to US Food and Drug Administration (FDA) Adverse Event Reporting System (FAERS). Specifically, we developed a support vector machine (SVM) model with recursive feature extraction, based on physicochemical and structural properties of compounds as model input. Cross validation demonstrates that the predictive model has a robust performance with averaged 70% of both sensitivity and specificity over 500 trials. An independent validation was performed on public benchmark drugs and the results suggest potential utility of our model for identifying safety alerts. This in silico approach, upon further validation, would ultimately be implemented, together with other in vitro safety assays, for screening compounds early in drug development. OPEN ACCESSCitation: Liu Y, Gao H, He YD (2020) A compound attributes-based predictive model for drug induced liver injury in humans. PLoS ONE 15(4): e0231252.

Detecting Novel Ototoxins and Potentiation of Ototoxicity by Disease Settings

“…Having a single target simplifies the model: what is the likelihood that the novel structure in question can serve as a ligand for hERG? By contrast, hepatoxicity comprises a heterogenous set of compound classes including protein kinase inhibitors, herbal supplements, and anti-cancer agents (157)(158)(159). Models of drug-induced liver injury predict phenotypic toxicity rather than specific ligand binding and therefore resemble ototoxicity in terms of the complexity needed for predictive power.…”

“…In this example, our model shares features with published ototoxicity models but we use a smaller but well-validated training dataset with strict classification of ototoxic vs. non-ototoxic drugs based on peer-reviewed literature drawn from PubMed. As input, we use isomorphic SMILES (simplified molecular-input line-entry system) data for each drug derived from PubChem; this format uses line notation to describe chemical structure as a series of fingerprints and is commonly used for toxicology modeling (151,159,162). This model uses a binary classifier approach where positive scores represent ototoxicity and negative scores represent non-ototoxicity.…”

Relationships between the Structure and Severe Drug-Induced Liver Injury for Low, Medium, and High Doses of Drugs

“…The prediction accuracy values (the invariant accuracy of prediction (IAP) 56 of 0.83 and balanced accuracy of 0.76) for the H-set of drugs is moderate but higher or comparable to those obtained in most of the published studies. 26,27,29,[31][32][33]35,36,39,40,43 The number of false-positive predictions is obviously higher among moderate-DILI-causing than that among non-DILI-causing drugs (Table 2).…”

“…In this regard, in silico methods offer a time- and cost-efficient way to estimate hepatotoxicity. The structure–activity relationships (SARs) allow predicting DILI based on the structural formula of a compound and, thus, are applicable at the early stages of drug development. − Plenty of SAR models for the prediction of DILI have been developed in recent years. − The accuracy of the models built varies from 0.6 to 0.9 depending on the data sets, methods, and descriptors. The ensemble approaches ,,, based on the integration of prediction results from many models and created using various machine learning techniques and descriptors, as well as those ,,,, based on deep learning, usually outperform “classical” ones based on the only machine learning method and the only type of descriptors.…”