Identification of structure-activity relationships for adverse effects of pharmaceuticals in humans: Part B. Use of (Q)SAR systems for early detection of drug-induced hepatobiliary and urinary tract toxicities

Mixed learning algorithms and features ensemble in hepatotoxicity prediction

“…For example, the three studies [4,24,25] which were proprietary in nature have large data sets (e.g. 1,266-1,608 compounds) but they were not disclosed and their models are unavailable or under licensing that restrict their distribution.…”

“…Therefore, a number of pure in silico hepatotoxicity prediction methods had been reported. These predictive models were generated from a variety of data sets of different endpoints related to hepatotoxicity and the models were made of different algorithms and methodologies [4,[22][23][24][25][26][27]. Two of these hepatotoxicity studies reported the use of consensus of support vector machine (SVM) or k-nearest neighbor (kNN) models trained from mixed training sets and optimized (mixed) features [26][27][28][29].…”

Applicability of QSAR analysis to the evaluation of the toxicological relevance of metabolites and degradates of pesticide active substances for dietary risk assessment

“…Matthews and co-workers (Matthews et al 2009b) have recently conducted an evaluation study to compare the performances of in-house models built using four QSAR software tools, CASE/MC4PC, MDL-QSAR, BioEpisteme, and Leadscope Predictive Data Miner), in predicting serious hepatobiliary and urinary tract toxicities of drugs Models were constructed for five types of liver injury (liver enzyme disorders, cytotoxic injury, cholestasis and jaundice, bile duct disorders, gall bladder disorders) and 6 types of urinary tract injury (acute renal disorders, nephropathies, bladder disorders, kidney function tests, blood in urine, urolithiases). The training set comprised approximately 1600 pharmaceuticals based on observations made in humans in pharmaceutical clinical trials and/or post-market surveillance by the FDA (Ursem et al 2009).…”

In silicomodels for drug-induced liver injury – current status

“…This trend is reflected in predictive models for DILI, most approaches being statistical using different methods (e.g. discriminant analysis [52], Bayesian models [65], artificial neural networks (ANN) [66], k-nearest neighbour quantitative structure--activity relationship (QSAR) [67,68], random forest (RF) [68] or QSAR software [69]), whereas only two published techniques involve developing structural alerts implemented in the knowledge-based expert systems (the Vertex cheminformatics platform (VERDI) [57] and Derek Nexus, formerly Derek for Windows [58]). …”