Search

Background: We address the problem of selecting and assessing classification and regression models using cross-validation. Current state-of-the-art methods can yield models with high variance, rendering them unsuitable for a number of practical applications including QSAR. In this paper we describe and evaluate best practices which improve reliability and increase confidence in selected models. A key operational component of the proposed methods is cloud computing which enables routine use of previously infeasible approaches., Methods: We describe in detail an algorithm for repeated grid-search V-fold cross-validation for parameter tuning in classification and regression, and we define a repeated nested cross-validation algorithm for model assessment. As regards variable selection and parameter tuning we define two algorithms (repeated grid-search cross-validation and double cross-validation), and provide arguments for using the repeated grid-search in the general case., Results: We show results of our algorithms on seven QSAR datasets. The variation of the prediction performance, which is the result of choosing different splits of the dataset in V-fold cross-validation, needs to be taken into account when selecting and assessing classification and regression models., Conclusions: We demonstrate the importance of repeating cross-validation when selecting an optimal model, as well as the importance of repeating nested cross-validation when assessing a prediction error.

In this article the Chemical Descriptors Library (CDL), a generic, open source software library for chemical informatics is introduced. The library is written using standard-compliant C++ programming language. The CDL provides a generic interface for traversing the structure of a molecular graph and accessing its properties. As a result, the software offers flexibility, reusability, and maintainability. This interface has been used to develop several chemical informatics algorithms, including molecular text format parsers and writers; substructure, pharmacophore, and atom type fingerprints; and both common substructure search and SMARTS search. The algorithms are described and evaluated on 3 data sets comprising 1000, 50000, and 100000 small molecules, respectively. The properties of the algorithms in terms of complexity analysis and processing times are presented and discussed.

As industrialization of laboratory processes for drug discovery continues to gather momentum, the bottleneck has moved toward exploitation of this tide of information to enable better quality decisions. The development of information-management systems to automate data and materials management can have a positive impact on productivity, as can increasingly sophisticated computer-aided molecular design approaches. However, as long as key decisions can only be taken by a small number of expert individuals working in a complex social environment, the impact of such innovations will be limited. This review describes Competitive Workflow, a distributed multi-agent system explicitly designed for the automation of decision making, currently the preserve of the expert. The approach builds on workflow architectures that capture best practice in information processing, but aims to extend these to model the tacit knowledge of the expert in the selection of alternative pathways through the workflow. The review also discusses recent developments in related workflow-management systems, particularly for information management and processing services front multiple sources, as well as distributed multi-agent approaches. A specific implementation of Competitive workflow--the Discovery Bus--and its application to meta-quantitative structure-activity relationship analysis is also described.

Early estimation of kinetics in man currently relies on extrapolation from experimental data generated in animals. Recent results from the application of a generic physiologically based model, Cloe PK) (Cyprotex), which is parameterised for human and rat physiology, to the estimation of plasma pharmacokinetics, are summarised in this paper. A comparison with predictive methods that involve scaling from in vivo animal data can also be made from recently published data. On average, the divergence of the predicted plasma concentrations from the observed data was 0.47 log units. For the external test set, > 70% of the predicted values of the AUC were within threefold of the observed values. Furthermore, the model was found to match or exceed the performance of three published interspecies scaling methods for estimating clearance, all of which showed a distinct bias towards overprediction. It is concluded that Cloe PK, as a means of integrating readily determined in vitro and/or in silico data, is a powerful, cost-effective tool for estimating exposure and kinetics in drug discovery and risk assessment that should, if widely adopted, lead to major reductions in the need for animal experimentation.

The routine assessment of xenobiotic in vivo kinetic behavior is currently dependent upon data obtained through animal experimentation, although in vitro surrogates for determining key absorption, distribution, metabolism, and elimination properties are available. Here we present a unique, generic, physiologically based pharmacokinetic (PBPK) model and demonstrate its application to the estimation of rat plasma pharmacokinetics, following intravenous dosing, from in vitro data alone. The model was parameterized through an optimization process, using a training set of in vivo data taken from the literature and validated using a separate test set of in vivo discovery compound data. On average, the vertical divergence of the predicted plasma concentrations from the observed data, on a semilog concentration-time plot, was approximately 0.5 log unit. Around 70% of all the predicted values of a standardized measure of area under the concentration-time curve (AUC) were within 3-fold of the observed values, as were over 90% of the training set t1/2 predictions and 60% of those for the test set; however, there was a tendency to overpredict t1/2 for the test set compounds. The capability of the model to rank compounds according to a given criterion was also assessed: of the 25% of the test set compounds ranked by the model as having the largest values for AUC, 61% were correctly identified. These validation results lead us to conclude that the generic PBPK model is potentially a powerful and cost-effective tool for predicting the mammalian pharmacokinetics of a wide range of organic compounds, from readily available in vitro inputs only.

Estimation of xenobiotic kinetics in humans frequently relies upon extrapolation from experimental data generated in animals. In an accompanying paper, we have presented a unique, generic, physiologically based pharmacokinetic model and described its application to the prediction of rat plasma pharmacokinetics from in vitro data alone. Here we demonstrate the application of the same model, parameterized for human physiology, to the estimation of plasma pharmacokinetics in humans and report a comparative evaluation against some recently published predictive methods that involve scaling from in vivo animal data. The model was parameterized through an optimization process, using a training set of in vivo data taken from the literature, and validated using a separate test set of published in vivo data. On average, the vertical divergence of the predicted plasma concentrations from the observed data, on a semilog concentration-time plot, was 0.47 log unit. For the training set, more than 80% of the predicted values of a standardized measure of the area under the concentration-time curve were within 3-fold of the observed values; over 70% of the test set predictions were within the same margin. Furthermore, in terms of predicting human clearance for the test set, the model was found to match or exceed the performance of three published interspecies scaling methods, all of which showed a distinct bias toward overprediction. We conclude that the generic physiologically based pharmacokinetic model, as a means of integrating readily determined in vitro and/or in silico data, is potentially a powerful, cost-effective tool for predicting human xenobiotic kinetics in drug discovery and risk assessment.

This paper describes a novel software architecture, Competitive Workflow, which implements workflow as a distributed and competitive multi-agent system. The implementation of a competitive workflow architecture designed to model important computer-aided molecular design workflows, the Discovery Bus, is described. QSPR modelling results for three example ADME datasets, for solubility, human plasma protein binding and P-glycoprotein substrates using an autonomous QSPR modelling workflow implemented on the Discovery Bus are presented. The autonomous QSPR system allows exhaustive exploration of descriptor and model space, automated model validation and continuous updating as new data and methods are made available. Prediction of properties of novel structures by an ensemble of models is also a feature of the system.

Purpose: To determine the extent of albumin diffusion from tissue pieces into medium during in vitro incubations, to develop and assess the utility of mathematical models describing this effect on the estimation of tissue-to-unbound plasma partition coefficients (Kpu) of drug substances and to derive factors to correct for associated errors., Methods: Twelve separate tissues were obtained from rats sacrificed by cervical dislocation, 48 h after an intravenous dose of 125I-human albumin, and tissue pieces incubated to determine the efflux of albumin into media over 2 to 4 h. A mathematical model was developed to predict and correct for the effect of albumin diffusion on the measured Kpu values of drugs., Results: The model predicted that the effect of albumin diffusion from tissue pieces during in vitro incubation (ranging from 14 to 59% remaining in tissue) on Kpu values was generally minimal, except for compounds that are highly plasma bound and have a low measured Kpu. Under these circumstances, the measured Kpu substantially underestimates the true value. An equation was derived from readily available or measurable parameters to correct for this underestimation., Conclusions: Albumin diffuses from tissue pieces into protein free media during in vitro incubations until equilibrium is reached, defined by the albumin Kpu. Model predictions indicated that for the majority of compounds albumin diffusion would have a minimal effect on the measured Kpu value and that a correction factor could be calculated to account for any deviation.

Purpose: To evaluate the ability to determine accurate in vivo tissue-to-unbound plasma distribution coefficients (Kpue) from in vitro data., Methods: Fresh pieces of fifteen rat tissues/organs were incubated at 37 degrees C with a homologous series of nine barbiturates covering a wide range of lipophilicity (Log P 0.02 to 4.13). Steady-state in vivo Kpue values were estimated from the tissue and plasma concentrations following simultaneous dosing by constant rate i.v. infusion of all nine barbiturates. Drug concentrations in the tissues and media were determined by HPLC with UV or mass spectrometric detection., Results: The pharmacokinetics of the barbiturate series following constant rate i.v. infusion indicated a range of clearance (0.49 to 30 ml x min(-1) x kg(-1)) and volume of distribution at steady state (0.51 to 1.9 l x kg(-1)) values. Good agreement was observed between the in vitro and in vivo Kpu values, although for the most lipophilic barbiturates the in vitro data underpredicted the in vivo tissue distribution for all tissues., Conclusion: The in vitro system for predicting the extent of in vivo tissue distribution works well for compounds of widely differing lipophilicity, although for the most lipophilic drugs it may result in an underprediction of in vivo values.

Simulation models for the prediction of pharmacokinetics in humans and other mammalian species, which are based on the physiology and mechanistic models of absorption, distribution, metabolism and elimination are reviewed. The structure of such models is explained with reference to papers describing the mathematical details and alternative representations of organ flow and distribution. Approaches to the modelling of more complex tissues such as tumours and the liver are also reviewed as well as some specific transport processes such as biliary secretion and methods of ADME property estimation by experimental and in silico models. Specific approaches to the modelling of gastro-intestinal transit are explained as is the extension of the approach to simulating drug-drug interactions following co-administration of more than one drug.

Purpose: The aim of this study was to evaluate the ability of an in vitro method of tissue distribution to accurately predict total water and extracellular aqueous spaces using marker compounds urea and inulin., Methods: Slices (50-200 mg) of all the major tissues in the rat were incubated with Hanks/HEPES pH7.4 buffer containing 14C-urea and 3H-inulin for 2 h at 37 degrees C. Tissue weight was noted before and after incubation and the tissue-to-buffer ratios determined., Results: 14C-Urea Kp estimates were generally greater than total tissue water due to tissue swelling, which varied widely among the tissues, up to 41% in muscle. In most cases, Kp values were much closer to in vivo values after correcting for the 14C-urea in the imbibed media (Kpcorr). The method was able to distinguish between 14C-urea and 3H-inulin Kp values and indicated that inulin occupied a smaller space than urea, which for the majority of tissues corresponded to the extracellular space., Conclusions: The Kp(corr) values for 14C-urea and Kp for 3H-inulin were consistent with total tissue water and extracellular space for the majority of tissues studied, indicating their suitability as marker compounds for checking the viability of this in vitro method for estimating tissue distribution.

Methods were developed to quantify a series of nine homologous 5-n-alkyl-5-ethyl barbituric acids in 15 rat tissues. Tissue homogenates were spiked with one of four multicomponent mixtures (methyl to n-propyl, n-propyl to n-pentyl, n-pentyl to n-heptyl and n-pentyl to n-nonyl). Liquid-liquid extraction was used to extract the homologues from the rat tissues. Reverse phase HPLC with UV detection at 214 nm was used to separate and quantify the individual barbiturates. The limit of detection for each respective homologue was 1 microg x g(-1) except skin and bone (2 microg x g(-1)). The methodology developed reduced a potential 135 individual assays to a more manageable 16.

Intestinal absorption of many hydrophilic drugs cannot be explained solely in terms of pH-partition and solvent-drag effects have been described in a number of cases. However, quantitative estimates of sieving coefficient (phi) for drug molecules have tended to be variable. In the present work an in situ perfused intestinal loop preparation in the rat has been used to measure the disappearance of five hydrophilic drugs from the intestinal lumen and a mathematical model of drug absorption in the presence of net and unidirectional fluid fluxes has been developed. The model allows separate estimation of the convective (solvent drag) and nonconvective (partition) components of drug absorption from the experimental data. The five drugs studied were found to have phi values ranging from 0.1-0.9; this was highly dependent on molecular size. Analysis of the data shows that three of the drugs are absorbed almost exclusively by the convective process (caffeine, cimetidine, hydrochlorthiazide) while the other two are absorbed by both convective and nonconvective processes (salicylate, oxprenolol). We conclude that the methodology is a useful and reliable means of deriving separate estimates of these two components of drug absorption.

Partition and distribution coefficients (log P, log D) into rat small intestinal brush border membrane (BBM) were measured for a variety of ionizable and non-ionizable drugs and solutes using a novel technique. The log P values were compared with those determined with model solvents, octanol and propylene glycol dipelargonate (PGDP). Non-ionizable solutes with log P values up to 3.0 showed that octanol was a better model for partition into the BBM than PGDP. With one exception, BBM partition coefficients of greater than 3 were not observed, even for solutes with log P values in model solvents that were greater than 5. Liposomes prepared from BBM lipids, or synthetic lipid mixtures of similar composition to BBM, demonstrated similar trends in partition coefficients to the intact BBM. Two cationic drugs, Atenolol and Xamoterol were investigated for partition into BBM lipid liposomes. An apparent enhancement of log D with respect to octanol was attributed to a "surfactant-like" orientation in the membrane and an interaction of the ionized drug with anionic phospholipid head groups. The anionic drug Proxicromil shows the expected decrease in log D with increasing pH, at low NaCl concentrations. Changes in electrophoretic mobility of liposomes after incorporation of Proxicromil into them were consistent with the negative charge of the ionized drug being at the membrane surface. It was concluded that Proxicromil also associates with membranes in a "surfactant-like" orientation and that increased extraction with increasing NaCl concentrations is a result of ionic strength effects. Partition of solutes into BBM vesicles is more complex than into organic solvents and probably represents an important step in overall intestinal permeation of solutes.

A new, calculated value of the van der Waals or intrinsic molecular volume VI is shown to be at least as effective as molar volume, V (the molecular weight divided by the liquid density), as a measure of the cavity term in linear solvation energy relationships for octanol-water partition coefficients and aqueous solubilities. Use of VI obviates the need for the empirical 10-mL/mol correction factor for aromatic and alicyclic solutes which was previously required, and which is shown here to arise from an underestimate of the cavity term due to reduced free volume in the pure liquid. In addition, since VI is a calculated quantity, equations which contain this term can be extended to compounds that are solids or gases in the pure state. Octanol-water partition coefficients, log P, of gases and solids are predicted accurately by the equation: log P = 0.41 + 5.14 VI/100 - 0.29 mu - 3.58 beta, where mu is the dipole moment and beta is the hydrogen bond acceptor basicity. Aqueous solubilities of some solids are reasonably well predicted by the equation: log Sw = 0.19 - 5.79 VI/100 + 0.24 mu + 4.95 beta - 0.01 (mp - 25), where mp is the melting point. The same equation without the melting point term gives good estimates of the comparable solubility of some gases.

The barriers to transport of methyl nicotinate across the water-octanol, water-2,2,4-trimethylpentane, and water-isopropyl myristate interfaces at 25 degrees C have been studied using a rotating diffusion cell. Comparison of results for systems which comprise zero, one, and two interfacial barriers shows that, contrary to previous reports, interfacial resistance is not a significant barrier in these cases and is below the limit of detection of the rotating diffusion cell method.