Prediction of ion channel activity using binary kernel discrimination

Voltage-gated ion channels are a diverse family of pharmaceutically important membrane proteins for which limited 3D information is available. A number of virtual screening tools have been used to assist with the discovery of new leads and with the analysis of screening results. One such tool, and the subject of this paper, is binary kernel discrimination (BKD), a machine-learning approach that has recently been applied to applications in chemoinformatics. It uses a training set of compounds, for which both structural and qualitative activity data are known, to produce a model that can then be used to rank another set of compounds in order of likely activity. Here, we report the use of BKD to build models for the prediction of five different ion channel targets using two types of activity data. The results obtained suggest that the approach provides an effective way of prioritizing compounds for acquisition and testing.     

Prediction of biological activity for high-throughput screening using binary kernel discrimination

High-throughput screening has made a significant impact on drug discovery, but there is an acknowledged need for quantitative methods to analyze screening results and predict the activity of further compounds. In this paper we introduce one such method, binary kernel discrimination, and investigate its performance on two datasets; the first is a set of 1650 monoamine oxidase inhibitors, and the second a set of 101 437 compounds from an in-house enzyme assay. We compare the performance of binary kernel discrimination with a simple procedure which we call "merged similarity search", and also with a feedforward neural network. Binary kernel discrimination is shown to perform robustly with varying quantities of training data and also in the presence of noisy data. We conclude by highlighting the importance of the judicious use of general pattern recognition techniques for compound selection.     

Virtual screening using binary kernel discrimination: analysis of pesticide data

This paper discusses the use of binary kernel discrimination (BKD) for identifying potential active compounds in lead-discovery programs. BKD was compared with established virtual screening methods in a series of experiments using pesticide data from the Syngenta corporate database. It was found to be superior to methods based on similarity searching and substructural analysis but inferior to a support vector machine. Similar conclusions resulted from application of the methods to a pesticide data set for which categorical activity data were available.     

A large descriptor set and a probabilistic kernel-based classifier significantly improve druglikeness classification

Probabilistic support vector machine (SVM) in combination with ECFP_4 (Extended Connectivity Fingerprints) were applied to establish a druglikeness filter for molecules. Here, the World Drug Index (WDI) and the Available Chemical Directory (ACD) were used as surrogates for druglike and nondruglike molecules, respectively. Compared with published methods using the same data sets, the classifier significantly improved the prediction accuracy, especially when using a larger data set of 341 601 compounds, which further pushed the correct classification rates up to 92.73%. On the other hand, most characteristic features for drugs and nondrugs found by the current method were visualized, which might be useful as guiding fragments for de novo drug design and fragment based drug design.     

A binary ant colony optimization classifier for molecular activities

Chemical fingerprints encode the presence or absence of molecular features and are available in many large databases. Using a variation of the Ant Colony Optimization (ACO) paradigm, we describe a binary classifier based on feature selection from fingerprints. We discuss the algorithm and possible cross-validation procedures. As a real-world example, we use our algorithm to analyze a Plasmodium falciparum inhibition assay and contrast its performance with other machine learning paradigms in use today (decision tree induction, random forests, support vector machines, artificial neural networks). Our algorithm matches established paradigms in predictive power, yet supplies the medicinal chemist and basic researcher with easily interpretable results. Furthermore, models generated with our paradigm are easy to implement and can complement virtual screenings by additionally exploiting the precalculated fingerprint information.     

Dynamic mixture probabilistic PCA classifier modeling and application for fault classification

A dynamic classifier based on the mixture probabilistic principal component analyzer (MPPCA) is proposed for fault classification. Compared with traditional methods, both fault detection and diagnosis are combined into a single classification task. By introducing a state indicator, the conventional MPPCA model is first designed as a standard classifier. Then, the static MPPCA model based classifier is temporally extended to the dynamic form within the hidden Markov model framework. Both static and dynamic MPPCA classifiers are obtained by using the Expectation‐Maximization algorithm. For performance evaluation, case studies of the continuous stirred tank heater process and the Tennessee Eastman process are carried out. Results indicate that the dynamic MPPCA classifier performs better compared with the static MPPCA classifier and the hidden Markov model based classifier. Copyright © 2015 John Wiley & Sons, Ltd.     

Simplistic classical probabilistic model of superexcited states of molecules

The decay processes in the superexcited state of a molecule are investigated theoretically in terms of a classical trajectory method. The simplest model of a diatomic molecule is considered. Particular attention is payed to the branching ratio for preionization and predissociation, and to the energy distribution of the ejected electrons. New formulas of practical use are derived for these two quantities.     

Virtual screening using binary kernel discrimination: effect of noisy training data and the optimization of performance

Binary kernel discrimination (BKD) uses a training set of compounds, for which structural and qualitative activity data are available, to produce a model that can then be applied to the structures of other compounds in order to predict their likely activity. Experiments with the MDL Drug Data Report database show that the optimal value of the smoothing parameter, and hence the predictive power of BKD, is crucially dependent on the number of false positives in the training set. It is also shown that the best results for BKD are achieved using one particular optimization method for the determination of the smoothing parameter that lies at the heart of the method and using the Jaccard/Tanimoto coefficient in the kernel function that is used to compute the similarity between a test set molecule and the members of the training set.     

Classifying the conformations of a chemical system using matrices of integers

A chemical system is a collection of atomic nuclei and accompanying electrons interacting by electromagnetic forces and organized into one or more molecules, ions, and transient structures. A general problem is to devise a mathematical representation of the conformation of a system which is useful in further mechanical analysis. Through dihedral rotations of bonds and relative motion of structures, many chemical systems can attain an infinite number of conformations. This paper is a theoretical presentation of a scheme to partition these conformations into a finite number of sets. The scheme involves a discrete aspect of chemical kinematics and uses matrices of integers called proximity matrices. The partitioning scheme is anticipated to be useful in studying reaction mechanisms and interactions between molecules, and in finding conformations of particular interest such as that with a potential energy near the global minimum for a molecule.     

Classifying large chemical data sets: using a regularized potential function method

In recent years classifiers generated with kernel-based methods, such as support vector machines (SVM), Gaussian processes (GP), regularization networks (RN), and binary kernel discrimination (BKD) have been very popular in chemoinformatics data analysis. Aizerman et al. were the first to introduce the notion of employing kernel-based classifiers in the area of pattern recognition. Their original scheme, which they termed the potential function method (PFM), can basically be viewed as a kernel-based perceptron procedure and arguably subsumes the modern kernel-based algorithms. PFM can be computationally much cheaper than modern kernel-based classifiers; furthermore, PFM is far simpler conceptually and easier to implement than the SVM, GP, and RN algorithms. Unfortunately, unlike, e.g., SVM, GP, and RN, PFM is not endowed with both theoretical guarantees and practical strategies to safeguard it against generating overfitting classifiers. This is, in our opinion, the reason why this simple and elegant method has not been taken up in chemoinformatics. In this paper we empirically address this drawback: while maintaining its simplicity, we demonstrate that PFM combined with a simple regularization scheme may yield binary classifiers that can be, in practice, as efficient as classifiers obtained by employing state-of-the-art kernel-based methods. Using a realistic classification example, the augmented PFM was used to generate binary classifiers. Using a large chemical data set, the generalization ability of PFM classifiers were then compared with the prediction power of Laplacian-modified naive Bayesian (LmNB), Winnow (WN), and SVM classifiers.     

The improvement of SIMCA classification by using kernel density estimation : Part 1. A new probabilistic classification technique and how to evaluate such a technique

One of the disadvantages of SIMCA pattern recognition is its inability to produce probabilistic classifications. Attempts to correct this involve distributional assumptions. It appears that SIMCA can handle the residual error terms efficiently, but that inside the class model subspace a crude truncation is used for determining a “normal range”, inside which all points are treated as equal. An improvement is made by applying kernel density estimation to the scores inside the class model subspace in combination with a normal error distribution in the remaining dimensions (CLASSY method). The evaluation of these probabilistic classification methods is discussed theoretically.     

Orientational order in binary mixtures of hard Gaussian overlap molecules

Based on a standard constant-pressure Monte Carlo molecular simulation, we have studied liquid crystal phases of binary mixtures of nonspherical molecules. The components of the mixtures are two types of hard Gaussian overlap (HGO) molecules. The first type of molecule has a small molecularelongation parameter (short HGO molecules) and cannot form stable liquid crystal phase in the bulk by themselves. The second type of molecule has a large elongation parameter (long HGO molecules) and can form a liquid crystal phase easily. In the mixtures, the short HGO molecules can form an orientationally ordered phase because the long HGO molecules form confining surfaces to induce the alignment of the short molecules. We also study the isotropic-nematic phase transition in different mixtures composed of short and long HGO molecules with different elongations and concentrations. The obtained result implies that small anisotropic molecules can show liquid crystal behavior.     

Direct correlation functions of binary mixtures of hard Gaussian overlap molecules

We study the direct correlation function (DCF) of a classical fluid mixture of nonspherical molecules. The components of the mixture are two types of hard ellipsoidal molecules with different elongations, interacting through the hard Gaussian overlap (HGO) model. Two different approaches are used to calculate the DCFs of this fluid, and the results are compared. Here, the Pynn approximation [J. Chem. Phys. 60, 4579 (1974)] is extended to calculate the DCF of the binary mixtures of HGO molecules, then we use a formalism based on the weighted density functional theory introduced by Chamoux and Perera [J. Chem. Phys. 104, 1493 (1996)]. These results are fairly in agreement with each other. The pressure of this system is also calculated using the Fourier zero components of the DCF. The results are in agreement with the Monte Carlo molecular simulation.     

Molecular dynamics simulation of structure H clathrate-hydrates of binary guest molecules

Molecular dynamics (MD) simulations are performed to study the stability of structure H clathrate-hydrates of methane+large-molecule guest substance (LMGS) at temperatures of 270, 273, 278 and 280 K under canonical (NVT-) ensemble condition in a 3×3×3 structure H unit cell replica with 918 TIP4P water molecules. The studied LMGS are 2-methylbutane (2-MB), 2,3-dimethylbutane (2,3-DMB), neohexane (NH), methylcyclohexane (MCH), adamantane and tert-butyl methyl ether (TBME). In the process of MD simulation, achieving equilibrium of the studied system is recognized by stability in calculated pressure for NVT-ensemble. So, for the accuracy of MD simulations, the obtained pressures are compared with the experimental phase diagrams. Therefore, the obtained equilibrium pressures by MD simulations are presented for studying the structure H clathrate-hydrates. The results show that the calculated temperature and pressure conditions by MD simulations are consistent with the experimental phase diagrams. Also, the radial distribution functions (RDFs) of host-host, host-guest and guest-guest molecules are used to analysis the characteristic configurations of the structure H clathrate-hydrate.     

A binary matrix for background suppression in MALDI-MS of small molecules

Application of matrix-assisted laser-desorption ionization mass spectrometry (MALDI-MS) to small-molecule detection is often limited, because of high matrix background signals in the low-mass region. We report here an approach in which a mixture of two conventional MALDI matrices with different proton affinity was used to suppress the formation of matrix clusters and fragments. Specifically, when acidic α-cyano-4-hydroxycinnamic acid (CHCA) and basic 9-aminoacridine (9-AA) were used as the binary matrix, fewer background matrix peaks were observed in both positive and negative-mode detection of small molecules. In addition, the presence of CHCA substantially reduced the laser fluence needed for analyte desorption and ionization; thus better signal-to-background ratios were observed for negatively charged inositol phosphates in complex plant extracts. The mixing of MALDI matrices of different protonaffinities leads to suppression of matrix clusterformation and subsequently yields cleaner MS spectraof fewer background peaks in both positive andnegative detection of small molecules     

Primary photoprocesses in molecules of carbocyanine dyes in binary solvent mixtures

The effect of the composition of the dimethyl sulfoxide (DMSO)-toluene mixture on the photophysical processes of thia-, indo-, and imidacarbocyanine dyes was studied. In the mixtures with the DMSO content of more than 20 vol %, the dyes representing solvated cations are characterized by high efficiency of trans → cis photoisomerization and fluorescence, in contrast to the extremely low efficiency of intersystem crossing to the triplet state. With growing the toluene content in the mixture, ion pairs between the dye cation and Cl⁻, Br⁻, I⁻, or BF⁴⁻ anion are formed. In the cases when the dye counterions are Br⁻ or I⁻, a sharp increase in the yield of the triplet molecules and a decrease in their lifetime take place. The results are discussed in terms of “the external heavy atom effect” in ion pairs.     

Thermal diffusion measurements and simulations of binary mixtures of spherical molecules

Thermal diffusion forced Rayleigh scattering measurements on binary mixtures of carbon tetrabromide (CBr(4)), tetraethylsilane, and di-tert-butylsilane in carbon tetrachloride (CCl(4)) are reported at different temperatures and concentrations. The Soret coefficient of CBr(4) in CCl(4) is positive and S(T) of both silanes in CCl(4) is negative, which implies that the heavier component always moves to the cold side. This is the expected behavior for unpolar simple molecules. Both silanes have the same mass so the influence of the difference in shape and moment of inertia could be studied. For all three systems, S(T) decreases with decreasing CCl(4) concentration. The results are discussed in the framework of thermodynamic theories and the Hildebrand parameter concept. Additionally, the Soret coefficients for both silaneCCl(4) systems were determined by nonequilibrium molecular-dynamics calculations. The simulations predict the correct direction of the thermophoretic motion and reflect the stronger drive toward the warm side for di-tert-butylsilane compared to the more symmetric tetraethylsilane. The values deviate systematically between 9% and 18% from the experimental values.