A Naive Bayes Classifier for Prediction of Multidrug Resistance Reversal Activity on the Basis of Atom Typing

Pharmacokinetics in drug discovery

The aim of this current review is to summarize the present status of pharmacokinetics in Drug Discovery. The review is structured into four sections. The first section is a general overview of what we understand by pharmacokinetics and the different LADMET aspects: Liberation, Absorption, Distribution, Metabolism, Excretion, and Toxicity. The second section highlights the different computational or in silico approaches to estimate/predict one or several aspects of the pharmacokinetic profile of a discovery lead compound. The third section discusses the most commonly used in vitro methodologies. The fourth and last section examines the various approaches employed towards the pharmacokinetic assessment of discovery molecules; including all the LADME processes, discussing the different mathematical methodologies available to establish the PK profile of a test compound; what the main differences are and what should be the criteria for using one or another mathematical approach. The major conclusion of this review is that the use of the appropriate preclinical assays has a key role in the long-term viability of a pharmaceutical company since applying the right tools early in discovery will play a key role in determining the company's ability to discover novel safe and effective therapeutics to patients as quickly as possible.

Discrimination of outer membrane proteins using machine learning algorithms

Discriminating outer membrane proteins (OMPs) from other folding types of globular and membrane proteins is an important task both for identifying OMPs from genomic sequences and for the successful prediction of their secondary and tertiary structures. In this work, we have analyzed the performance of different methods, based on Bayes rules, logistic functions, neural networks, support vector machines, decision trees, etc. for discriminating OMPs. We found that most of the machine learning techniques discriminate OMPs with similar accuracy. The neural network-based method could discriminate the OMPs from other proteins [globular/transmembrane helical (TMH)] at the fivefold cross-validation accuracy of 91.0% in a dataset of 1,088 proteins. The accuracy of discriminating globular proteins is 88.8% and that of TMH proteins is 93.7%. Further, the neural network method is tested with globular proteins belonging to 30 different folding types and it could successfully exclude 95% of the considered proteins. The proteins with SAM domain such as knottins, rubredoxin, and thioredoxin folds are eliminated with 100% accuracy. These accuracy levels are comparable to or better than other methods in the literature. We suggest that this method could be effectively used to discriminate OMPs and for detecting OMPs in genomic sequences.

An Accurate and Interpretable Bayesian Classification Model for Prediction of hERG Liability

Drug-induced QT interval prolongation has been identified as a critical side-effect of non-cardiovascular therapeutic agents and has resulted in the withdrawal of many drugs from the market. As almost all cases of drug-induced QT prolongation can be traced to the blockade of a voltage-dependent potassium ion channel encoded by the hERG (the human ether-à-go-go-related gene), early identification of potential hERG channel blockers will decrease the risk of cardiotoxicity-induced attritions in the later and more expensive development stage. Presented herein is a naive Bayes classifier to categorize hERG blockers into active and inactive classes, by using a universal, generic molecular descriptor system.1 The naive Bayes classifier was built from a training set containing 1979 corporate compounds, and exhibited an ROC accuracy of 0.87. The model was validated on an external test set of 66 drugs, of which 58 were correctly classified. The cumulative probabilities reflected the confidence of prediction and were proven useful for the identification of hERG blockers. Relative performance was compared for two classifiers constructed from either an atom-type-based molecular descriptor or the long range functional class fingerprint descriptor FCFP_6. The combination of an atom-typing descriptor and the naive Bayes classification technique enables the interpretation of the resulting model, which offers extra information for the design of compounds free of undesirable hERG activity.