Quantitative Structure-activity Relationship Field Research Articles

Study of PARP inhibitors for breast cancer based on enhanced multiple kernel function SVR with PSO.

PARP1 is one of six enzymes required for the highly error-prone DNA repair pathway microhomology-mediated end joining (MMEJ) and needs to be inhibited when over-expressed. In order to study the PARP1 inhibitory effect of fused tetracyclic or pentacyclic dihydrodiazepinoindolone derivatives (FTPDDs) by quantitative structure-activity relationship technique, six models were established by four kinds of methods, heuristic method, gene expression programming, random forester, and support vector regression with single, double, and triple kernel function respectively. The single, double, and triple kernel functions were RBF kernel function, the integration of RBF and polynomial kernel functions, and the integration of RBF, polynomial, and linear kernel functions respectively. The problem of multi-parameter optimization introduced in the support vector regression model was solved by the particle swarm optimization algorithm. Among the models, the model established by support vector regression with triple kernel function, in which the optimal R 2 and RMSE of training set and test set were 0.9353, 0.9348 and 0.0157, 0.0288, and R2 cv of training set and test set were 0.9090 and 0.8971, shows the strongest prediction ability and robustness. The method of support vector regression with triple kernel function is a great promotion in the field of quantitative structure-activity relationship, which will contribute a lot to designing and screening new drug molecules. The information contained in the model can provide important factors that guide drug design. Based on these factors, six new FTPDDs have been designed. Using molecular docking experiments to determine the properties of new derivatives, the new drug was ultimately successfully designed.

The VEGAHUB Platform: The Philosophy and the Tools.

VEGAHUB (www.vegahub.eu) is a repository of freely available, downloadable tools based on computational toxicology methodologies. The main software tool available in VEGAHUB is VEGA QSAR software encoding more than 90 quantitative structure-activity relationship (QSAR) models for tens of endpoints for human toxicology, ecotoxicology, environmental, physico-chemical and toxicokinetic properties. However, beyond VEGA QSAR, VEGAHUB offers several other tools. Here, we present these resources, the possibilities to fully exploit them and the ways in which to integrate results provided by different VEGAHUB tools. Read-across and weight-of-evidence represent a major advantage of VEGAHUB. Integration between hazard and exposure is provided within innovative tools, which are specific for well-defined scenarios, such as those for cosmetic products. Prioritisation can be achieved by integrating results from 48 models. Finally, we highlight how some tools may not only fit predefined endpoints but also could be applied to general problems and research applications in the QSAR field. A couple of examples are provided, in which a critical assessment of the predictions and the documentation associated with the prediction are considered, in order to properly assess the quality of the results. These results may be associated with different levels of uncertainty or even be conflicting.

The effect of noise on the predictive limit of QSAR models

A key challenge in the field of Quantitative Structure Activity Relationships (QSAR) is how to effectively treat experimental error in the training and evaluation of computational models. It is often assumed in the field of QSAR that models cannot produce predictions which are more accurate than their training data. Additionally, it is implicitly assumed, by necessity, that data points in test sets or validation sets do not contain error, and that each data point is a population mean. This work proposes the hypothesis that QSAR models can make predictions which are more accurate than their training data and that the error-free test set assumption leads to a significant misevaluation of model performance. This work used 8 datasets with six different common QSAR endpoints, because different endpoints should have different amounts of experimental error associated with varying complexity of the measurements. Up to 15 levels of simulated Gaussian distributed random error was added to the datasets, and models were built on the error laden datasets using five different algorithms. The models were trained on the error laden data, evaluated on error-laden test sets, and evaluated on error-free test sets. The results show that for each level of added error, the RMSE for evaluation on the error free test sets was always better. The results support the hypothesis that, at least under the conditions of Gaussian distributed random error, QSAR models can make predictions which are more accurate than their training data, and that the evaluation of models on error laden test and validation sets may give a flawed measure of model performance. These results have implications for how QSAR models are evaluated, especially for disciplines where experimental error is very large, such as in computational toxicology.Graphical

DeepSnap-Deep Learning Approach Predicts Progesterone Receptor Antagonist Activity With High Performance.

The progesterone receptor (PR) is important therapeutic target for many malignancies and endocrine disorders due to its role in controlling ovulation and pregnancy via the reproductive cycle. Therefore, the modulation of PR activity using its agonists and antagonists is receiving increasing interest as novel treatment strategy. However, clinical trials using the PR modulators have not yet been found conclusive evidences. Recently, increasing evidence from several fields shows that the classification of chemical compounds, including agonists and antagonists, can be done with recent improvements in deep learning (DL) using deep neural network. Therefore, we recently proposed a novel DL-based quantitative structure-activity relationship (QSAR) strategy using transfer learning to build prediction models for agonists and antagonists. By employing this novel approach, referred as DeepSnap-DL method, which uses images captured from 3-dimension (3D) chemical structure with multiple angles as input data into the DL classification, we constructed prediction models of the PR antagonists in this study. Here, the DeepSnap-DL method showed a high performance prediction of the PR antagonists by optimization of some parameters and image adjustment from 3D-structures. Furthermore, comparison of the prediction models from this approach with conventional machine learnings (MLs) indicated the DeepSnap-DL method outperformed these MLs. Therefore, the models predicted by DeepSnap-DL would be powerful tool for not only QSAR field in predicting physiological and agonist/antagonist activities, toxicity, and molecular bindings; but also for identifying biological or pathological phenomena.

A New Model of Atomic Nucleophilicity Index and Its Application in the Field of QSAR

A new ansatz is suggested for computing the atomic nucleophilicity index (N) for atoms of 103 elements of the periodic table resting upon the mutual action of two periodic properties, atomic polarizability (α) and effective nuclear charge (Zeff). The effectiveness of the model is illustrated by the explicit periodic behaviour. In addition, molecular nucleophilicity (NAM) is being proposed as an arithmetic mean of the atomic nucleophilicities of the constituent atoms of the given molecule. Due to the nonexistence of a benchmark for atomic nucleophilicity, molecular nucleophilicity index is evaluated and a comparative analysis is made with the existing data as a validity test. Furthermore, computed density functional theory (DFT) based reactivity descriptor, viz. atomic nucleophilicity index, have been employed to construct a quantitative structure–activity relationship (QSAR) model, using regression analysis, to study the biological activities of testosterone derivatives.

PyDescriptor: A new PyMOL plugin for calculating thousands of easily understandable molecular descriptors

The field of Quantitative Structure-Activity Relationship (QSAR) relies heavily on molecular descriptors. Among various guidelines suggested by Organisation for Economic Co-operation and Development (OECD), a very important guideline demands the mechanistic interpretation of a QSAR model. For this, a very attractive idea is to build a QSAR model using easily understandable molecular descriptors. To address this important issue, in the present work, we present an innovative chem-informatics tool, PyDescriptor. It can calculate a diverse pool of 11,145 molecular descriptors comprising easily understandable 1D- to 3D- descriptors encoding pharmacophoric patterns, atomic fragments and a variety of fingerprints. It is a new Python based plugin implemented within the commonly used visualization software PyMOL. PyDescriptor has several advantages like easy to install, open source, works on all major platforms (Windows, Linux, MacOS), easy to use through graphical user interface (GUI) and command-line, and output is saved in comma separated values (CSV) file format for further QSAR procedure. The plugin is freely available for academia.

Recent Advances in Multi-Task QSAR Modeling for Drug Design

Computer-aided drug design may provide a massive scope to reduce the overall duration as well as the expenditure associated with drug discovery. Although the concept of Quantitative Structure Activity Relationship (QSAR) is almost fifty years old, it is still considered as a very useful method for finding novel therapeutic agents. In the last decade, the field of QSAR has witnessed great advances in terms of techniques, ideas and approaches. 3 Multi-task QSAR is one such approach that has a huge opportunity to contribute in the discovery of novel therapeutic agents. Classical QSAR techniques focus on the classical idea that one drug is aimed to act on one target. However, due to the highly complex and multi-factorial nature of some diseases like cancer, microbial infections, neurodegenerative and metabolic disorders, the classical idea of drug development frequently produces less satisfying effects. The effect of the inhibition of one specific target is compensated by different back-up or fail-safe mechanisms. Polypharmacology focuses on the concept of simultaneous inhibition of multiple targets by single therapeutic agents. Contrary to single target drugs that bind with a specific target with high affinity, multi-target agents are low affinity binders. Therefore, management of therapy is easier for the multi-target agents. Multi-task QSAR techniques are able to incorporate the ideas of polypharmacology. While conventional QSAR techniques are based on the datasets of compounds with a single target and same or similar assay conditions, multi-task QSAR can consider data on multiple targets and various assay conditions at a time. In addition, multi-task QSAR techniques are also useful to address the problems of fewer datapoints. QSAR is a statistical technique and a higher amount of data-points generates more reliable models. However, for several diseases the availability of a data is less and developed models on the available data may not be able to provide detailed insight into structural requirements of the molecules for higher potency. As multi-task QSAR models may be developed on the data collected from different assay conditions, more information may be retrieved for the explanation of structural patterns. Therefore, multi-task QSAR has emerged as one of the most essential tools in the last few years and various groups of researchers participated in the development of multi-task QSAR models. Though conventional 2D QSAR techniques have found several alternatives such as 3D QSAR, pharmacophore mapping as well as molecular docking, 2D-QSAR is unique in handling problems of multi-task modeling. Multi-task QSAR has been utilized for the developing models for anti-cancer, anti-bacterial 17 and neuroprotective agents. In addition, reports on multi-target modeling on kinase and ubiquitineproteosome pathways are also found. Furthermore, multi-task agents are also useful for understanding structural patterns required for toxicities of nanoparticles, environmental pollutants etc. Although development of multi-task QSAR model is quite similar to conventional modeling as far as descriptor calculation and model generation are concerned, the construction and management of a dataset is quite different. In most of the cases, classification models are generated with the data by suitable machine learning tools such as Linear Discriminant Analyses (LDA), Artificial Neural Network (ANN), Support Vector Machine (SVM) etc. Generally, a specific cut-off value for each system is fixed by the user to classify the data into ‘sensitive, and ‘non-sensitive’. Sometimes special statistical treatments, like averaging the descriptor values for specific system and calculation of moving average, are required for handling the complicated data. Some attempts were taken to develop regression based multi-target models. However, there still is a debate about the actual effectiveness of multi-target QSAR modeling as the size of a data set is sometimes so increased that the model may not be properly validated by conventional approaches. Moreover, there are concerns over applicability domain of the multitask models. Nevertheless, multi-task QSAR models may provide opportunity to widen the area of QSAR studies as well as to contribute to the investigation of different emerging areas such as nanotoxicity, polypharmacology, drug repositioning, environmental safety management etc.

Cheminformatics and Molecular Modeling

More than fifty years have passed since the field of Quantitative Structure-Activity Relationships (QSAR) modeling was founded by Corwin Hansch [1]. Initially conceptualized as the logical extension of physical organic chemistry, QSAR modeling has grown, diversified, and evolved from application to small series of congeneric compounds using relatively simple regression methods to the analysis of very large datasets comprising thousands of diverse molecular structures using a wide variety of statistical and machine learning techniques. More than fifty years of continuous improvements, interdisciplinary breakthroughs, and community driven developments were needed to make QSAR one of the commonly employed approaches to modeling the physical and biological properties of chemicals in use today.

Exploring the activity space of peptides binding to diverse SH3 domains using principal property descriptors derived from amino acid rotamers

Although there were intensive works addressed on multivariate extraction of the informative components from numerous physicochemical parameters of amino acids in isolated state, the various conformational behaviors of amino acids in complicated biological context have long been underappreciated in the field of quantitative structure-activity relationship (QSAR). In this work, the amino acid rotamers, which were derived from statistical survey of protein crystal structures, were used to reproduce the conformational variety of amino acid side-chains in real condition. In this procedure, these rotamers were superposed into a nx x ny x nz lattice and an artificial probe was employed to detect four kinds of nonbonding field potentials (i.e., electrostatic, steric, hydrophobic, and hydrogen bonds) at each lattice point using a Gaussian-type potential function; the generated massive data were then subjected to a principal component analysis (PCA) treatment to obtain a set of few, informative amino acid descriptors. We used this set of descriptors, that we named principal property descriptors derived from amino acid rotamers (PDAR), to characterize over 13,000 peptides with known binding affinities to 10 types of SH3 domains. Genetic algorithm/ partial least square regression (GA/PLS) modeling and Monte Carlo cross-validation (MCCV) demonstrated that the correlation between the PDAR descriptors and the binding affinities of peptides are comparable with or even better than previously published models. Furthermore, from the PDAR-based QSAR models we concluded that the core motif of peptides, particularly the electrostatic property, hydrophobicity, and hydrogen bond at residue positions P3, P2, and/or P0, contribute significantly to the hAmph SH3 domain-peptide binding, whereas two ends of the peptides, such as P6, P4, P-4, and P5, only play a secondary role in the binding.

Optimal assignment methods for ligand-based virtual screening

BackgroundLigand-based virtual screening experiments are an important task in the early drug discovery stage. An ambitious aim in each experiment is to disclose active structures based on new scaffolds. To perform these "scaffold-hoppings" for individual problems and targets, a plethora of different similarity methods based on diverse techniques were published in the last years. The optimal assignment approach on molecular graphs, a successful method in the field of quantitative structure-activity relationships, has not been tested as a ligand-based virtual screening method so far.ResultsWe evaluated two already published and two new optimal assignment methods on various data sets. To emphasize the "scaffold-hopping" ability, we used the information of chemotype clustering analyses in our evaluation metrics. Comparisons with literature results show an improved early recognition performance and comparable results over the complete data set. A new method based on two different assignment steps shows an increased "scaffold-hopping" behavior together with a good early recognition performance.ConclusionThe presented methods show a good combination of chemotype discovery and enrichment of active structures. Additionally, the optimal assignment on molecular graphs has the advantage to investigate and interpret the mappings, allowing precise modifications of internal parameters of the similarity measure for specific targets. All methods have low computation times which make them applicable to screen large data sets.

Gaussian process: an alternative approach for QSAM modeling of peptides

Different statistical modeling methods (SMMs) are used for nonlinear system classification and regression. On the basis of Bayesian probabilistic inference, Gaussian process (GP) is preliminarily used in the field of quantitative structure-activity relationship (QSAR) but has not yet been applied to quantitative sequence-activity model (QSAM) of biosystems. This paper proposes the application of GP as an alternative tool for the QSAM modeling of peptides. To investigate the modeling performance of GP, three classical peptide panels were used: Angiotensin-I converting enzyme inhibitory dipeptides, bradykinin-potentiating pentapeptides and cationic antimicrobial pentadecapeptides. On this basis, we made a comprehensive comparison between the GP and some widely used SMMs such as PLS, artificial neural network (ANN) and support vector machine (SVM), and gave the conclusions as follow: (1) for those of structurally complicated peptides, particularly the polypeptides, linear PLS was incapable of capturing all dependences hidden in the peptide systems, (2) even in assistance with the monitoring technique, ANN was inclined to be overtrained in the cases of insufficient number of peptide samples, (3) SVM and GP performed best for the three peptide panels. Moreover, since GP was able to correlate the linear and nonlinear-hybrid relationship, it was slightly superior to SVM at most peptide sets.

The proposal of architecture for chemical splitting to optimize QSAR models for aquatic toxicity

One of the challenges in the field of quantitative structure-activity relationship (QSAR) analysis is the correct classification of a chemical compound to an appropriate model for the prediction of activity. Thus, in previous studies, compounds have been divided into distinct groups according to their mode of action or chemical class. In the current study, theoretical molecular descriptors were used to divide 568 organic substances into subsets with toxicity measured for the 96-h lethal median concentration for the Fathead minnow ( Pimephales promelas). Simple constitutional descriptors such as the number of aliphatic and aromatic rings and a quantum chemical descriptor, maximum bond order of a carbon atom divide compounds into nine subsets. For each subset of compounds the automatic forward selection of descriptors was applied to construct QSAR models. Significant correlations were achieved for each subset of chemicals and all models were validated with the leave-one-out internal validation procedure ( R cv 2 ≈ 0.80 ) . The results encourage to consider this alternative way for the prediction of toxicity using QSAR subset models without direct reference to the mechanism of toxic action or the traditional chemical classification.

Spatial Sign Preprocessing: A Simple Way To Impart Moderate Robustness to Multivariate Estimators

The spatial sign is a multivariate extension of the concept of sign. Recently multivariate estimators of covariance structures based on spatial signs have been examined by various authors. These new estimators are found to be robust to outlying observations. From a computational point of view, estimators based on spatial sign are very easy to implement as they boil down to a transformation of the data to their spatial signs, from which the classical estimator is then computed. Hence, one can also consider the transformation to spatial signs to be a preprocessing technique, which ensures that the calibration procedure as a whole is robust. In this paper, we examine the special case of spatial sign preprocessing in combination with partial least squares regression as the latter technique is frequently applied in the context of chemical data analysis. In a simulation study, we compare the performance of the spatial sign transformation to nontransformed data as well as to two robust counterparts of partial least squares regression. It turns out that the spatial sign transform is fairly efficient but has some undesirable bias properties. The method is applied to a recently published data set in the field of quantitative structure-activity relationships, where it is seen to perform equally well as the previously described best linear model for these data.

3D QSAR in modern drug design.

The belief that there is a direct relationship between chemical structure and biological activity of therapeutic agents is fundamental to the field of medicinal chemistry. Indeed, the efforts of early medicinal chemists focused on well-defined structural modifications to active lead compounds as molecules were developed into drugs. The relationship between chemical properties like solvent partitioning and biological activity was recognized over a century ago [1]. Almost 40 years ago, Hansch, Fujita and co-workers invented the field of Quantitative Structure-Activity Relationships (QSAR) [2]. In this approach hole molecule parameters such as LogPo/w (the partition coefficient for 1-octanol/water partitioning),[3] molar refractivity, shape and topology indices [4], etc. for groups of related compounds are statistically correlated with measures of biological activity to obtain a QSAR equation. This equation relates easy to measure (or predict/calculate) values for molecules to the more difficult to measure biological activities. Once a QSAR is obtained, verified and found to be “predictive,” the biological activity for new chemical entities that may or may not exist can easily be predicted. Numerous success stories from QSAR over the past four decades validate the fundamental relationship between structure and activity.

Review Biologicals & Immunologicals: Applications of artificial neural networks to quantitative structure-activity relationships

his review article examines the role of supervised and unsupervised artificial neural networks (ANNs) in the field of quantitative structure-activity relationships (QSAR). They were found to be useful in both classical QSAR and structure-property correlation (SPC) studies where the data sets contain significant non-linear relationships. New applications of ANNs, including methods of descriptor optimisation, simultaneous prediction of multiple descriptors, the development of flexible pharmacophore models and new methods for the multidimensional reduction and display of data sets suggest that ANNs will have a useful role in the field of QSAR in the future.

New techniques enhance structure activity work

A group of industrial chemists and several academic collaborators are fomenting revolution in the normally sedate field of quantitative structureactivity relationships (QSAR). They are developing a series of QSAR techniques they hope will replace some traditional QSAR concepts that they believe are over the hill. In QSAR, mathematical descriptors of molecular structure are used as a basis for predicting chemical and physical properties or biological activity. QSAR is widely used for drug discovery, among other applications. For example, it can be employed to predict partition coefficients, lipophilicities, or relative reactivities with biological receptors to support drug-design efforts. But Jack W. Frazer of Tuolumne, Calif., team leader of a scientific computing group that consults for Eastman Kodak, believes QSAR techniques up to now have not realized their true potential. Several years ago, Frazer's group was asked to bring new insights and computer technology to the drug discovery process at Koda...