Different Sets Of Descriptors Research Articles

Development of a robust Machine learning model for Ames test outcome prediction

The mutagenicity is an essential parameter for evaluating the safety of pharmaceuticals, chemicals, consumer products, environmentally related compounds and the Ames assay is a significant test for predicting the mutagenicity of chemical compounds. In the data-driven era, developing robust models for efficient mutagenicity prediction before synthesizing and testing in vitro has gained increasing attention. In this study, a machine learning model that could predict Ames mutagenicity based on 2D molecular descriptors was developed. A multistep filtering process that adequately helps in identifying significant descriptors was adopted in this study. Three different sets of descriptors, namely, RDKit, Mordred and CDK were used to train three machine learning algorithms, viz., random forest, xgboost and catboost. The datasets were collected from different resources to develop a robust machine learning model. The robustness of this model was further validated by comparing different available ML and DL models for Ames genotoxicity. Specifically, 12 models, including our xgboost model, were used to validate an external dataset, and our model exhibited excellent performance, with an impressive AUC of 0.97. The codes to predict the genotoxicity of a molecule is available at https://github.com/Naga270588/Genotoxicity.

Effect of Data Quality and Data Quantity on the Estimation of Intrinsic Solubility: Analysis Based on a Single-Source Data Set.

Aqueous solubility is one of the most important physicochemical properties of drug molecules and a major driving force for oral drug absorption. To date, the performance of in silico models for the estimation of solubility for novel chemical space is limited. To investigate possible reasons and remedies for this, the Johnson and Johnson in-house aqueous solubility data with over 40,000 compounds was leveraged. All data were generated through the same high-throughput assay, providing a unique opportunity to explore the relationship between data quality, quantity, and model estimations. Six intrinsic solubility data sets with different sizes and noise levels were generated by making use of three different approaches: (i) inclusion or exclusion of amorphous solid residue, (ii) measured or experimental log D to identify the intrinsic solubility, and (iii) adopting or omitting a quality check process in the data processing workflow. A random forest regressor was trained on the data sets with three different sets of descriptors calculated from RDKit, ADMET predictor, or Mordred, and the performances were evaluated with nested cross-validation as well as ten refined test sets. The models confirm, as expected, that with the same data set size, high-quality data leads to better model performance; however, also, models trained with larger data sets containing analytical variability can give equally accurate estimations compared to models trained with small, clean, and diverse data sets. However, noise introduced by including the presence of amorphous solid postsolubility measurement in the training data set cannot be overcome by increasing data size, as they are introducing a biased systematic positive error in the data set, confirming the importance of critical data review. Finally, two top-performing models were tested on the first test set from the second solubility challenge, achieving RMSE values of 0.74 and 0.72 and log S ± 0.5 of 46 and 48%, respectively. These results demonstrated improved performance compared to those reported in the findings of the competition, highlighting that a single-source curated data set can enhance the prediction of intrinsic solubility.

A QSAR study for predicting malformation in zebrafish embryo

Background Developmental toxicity tests are extremely expensive, require a large number of animals, and are time-consuming. It is necessary to develop a new approach to simplify the analysis of developmental endpoints. One of these endpoints is malformation, and one group of ongoing methods for simplifying is in silico models. In this study, we aim to develop a quantitative structure–activity relationship (QSAR) model and identify the best algorithm for predicting malformations, as well as the most important and effective physicochemical properties associated with malformation. Methods The dataset was extracted from a reliable database called COMPTOX. Physicochemical properties (descriptors) were calculated using Mordred and RDKit chemoinformatics software. The data were cleaned, preprocessed, and then split into training and testing sets. Machine learning algorithms, such as gradient boosting model (GBM) and logistic regression (LR), as well as deep learning models, including multilayer perceptron (MLP) and neural networks (NNs) trained with train set data and different sets of descriptors. The models were then validated with test set and various statistical parameters, such as Matthew’s correlation coefficient (MCC) and balanced accuracy (BAC) score, were used to compare the models. Results A set of descriptors containing with 78% AUC was identified as the best set of descriptors. Gradient boosting was determined to be the best algorithm with 78% predictive power. Conclusions The descriptors that were the most effective for developing models directly impact the mechanism of malformation, and GBM is the best model due to its MCC and BAC.

How Granular Can a Dose Form Be Described? Considering EDQM Standard Terms for a Global Terminology

The aim was (1) to analyse the features of the EDQM terminology, (2) to formulate proposals for minor changes and (3) to create a small ontology of dose forms, based on characteristics of EDQM, and suitable for alignment with other dose form terminologies. The 428 Pharmaceutical Dose Forms (PDF) (“human and veterinary” only) were extracted from the EDQM Standard Terms database. A quantitative and qualitative analysis of the textual definitions of the terms was conducted. Through an analysis of unique combinations of different sets of descriptors and characteristics, a small ontology was built in three levels. For the 143 transformable PDFs, the administrable dose form was made explicit, with 121 requiring only one transformation and 22 multiple transformations, of which 10 include “no transformation”. Different levels of aggregations of the 428 PDFs were tested in 4 analyses, ranging from 206 to 383 unique combinations. An ontology in Webprotégé was created of 22 higher-level concepts (based on the intended site characteristics) and 69 intermediate-level terms (newly created) to accommodate the 428 PDFs of EDQM. EDQM Dose Form terminology is suitable terminology in terms of granularity, for defining dose forms of medicinal products, to enable fair comparison of similar medicinal products, and global identification of medicinal products (IDMP). Recommendations for minor improvements and a simple ontology for dose forms are proposed.

The role of feature space in atomistic learning

Efficient, physically-inspired descriptors of the structure and composition of molecules and materials play a key role in the application of machine-learning techniques to atomistic simulations. The proliferation of approaches, as well as the fact that each choice of features can lead to very different behavior depending on how they are used, e.g. by introducing non-linear kernels and non-Euclidean metrics to manipulate them, makes it difficult to objectively compare different methods, and to address fundamental questions on how one feature space is related to another. In this work we introduce a framework to compare different sets of descriptors, and different ways of transforming them by means of metrics and kernels, in terms of the structure of the feature space that they induce. We define diagnostic tools to determine whether alternative feature spaces contain equivalent amounts of information, and whether the common information is substantially distorted when going from one feature space to another. We compare, in particular, representations that are built in terms of n-body correlations of the atom density, quantitatively assessing the information loss associated with the use of low-order features. We also investigate the impact of different choices of basis functions and hyperparameters of the widely used SOAP and Behler–Parrinello features, and investigate how the use of non-linear kernels, and of a Wasserstein-type metric, change the structure of the feature space in comparison to a simpler linear feature space.

Minimal-uncertainty prediction of general drug-likeness based on Bayesian neural networks

Triaging unpromising lead molecules early in the drug discovery process is essential for accelerating its pace while avoiding the costs of unwarranted biological and clinical testing. Accordingly, medicinal chemists have been trying for decades to develop metrics—ranging from heuristic measures to machine-learning models—that could rapidly distinguish potential drugs from small molecules that lack drug-like features. However, none of these metrics has gained universal acceptance and the very idea of ‘drug-likeness’ has recently been put into question. Here, we evaluate drug-likeness using different sets of descriptors and different state-of-the-art classifiers, reaching an out-of-sample accuracy of 87–88%. Remarkably, because these individual classifiers yield different Bayesian error distributions, their combination and selection of minimal-variance predictions can increase the accuracy of distinguishing drug-like from non-drug-like molecules to 93%. Because total variance is comparable with its aleatoric contribution reflecting irreducible error inherent to the dataset (as opposed to the epistemic contribution due to the model itself), this level of accuracy is probably the upper limit achievable with the currently known collection of drugs. When designing new drugs, there are countless ways to create molecules, yet only a few interact with biological targets. Beker and colleagues provide here a graph neural network based metric for drug-likeness that can guide the search.

A Quantitative Model for Alkane Nucleophilicity Based on C-H Bond Structural/Topological Descriptors.

A first quantitative model for calculating the nucleophilicity of alkanes is described. A statistical treatment was applied to the analysis of the reactivity of 29 different alkane C-H bonds towards in situ generated metal carbene electrophiles. The correlation of the recently reported experimental reactivity with two different sets of descriptors comprising a total of 86 parameters was studied, resulting in the quantitative descriptor-based alkane nucleophilicity (QDEAN) model. This model consists of an equation with only six structural/topological descriptors, and reproduces the relative reactivity of the alkane C-H bonds. This reactivity can be calculated from parameters emerging from the schematic drawing of the alkane and a simple set of sums.

A Quantitative Model for Alkane Nucleophilicity Based on C−H Bond Structural/Topological Descriptors

AbstractA first quantitative model for calculating the nucleophilicity of alkanes is described. A statistical treatment was applied to the analysis of the reactivity of 29 different alkane C−H bonds towards in situ generated metal carbene electrophiles. The correlation of the recently reported experimental reactivity with two different sets of descriptors comprising a total of 86 parameters was studied, resulting in the quantitative descriptor‐based alkane nucleophilicity (QDEAN) model. This model consists of an equation with only six structural/topological descriptors, and reproduces the relative reactivity of the alkane C−H bonds. This reactivity can be calculated from parameters emerging from the schematic drawing of the alkane and a simple set of sums.

Instanton rate constant calculations using interpolated potential energy surfaces in nonredundant, rotationally and translationally invariant coordinates.

A trivial flaw in the utilization of artificial neural networks in interpolating chemical potential energy surfaces (PES) whose descriptors are Cartesian coordinates is their dependence on simple translations and rotations of the molecule under consideration. A different set of descriptors can be chosen to circumvent this problem, internuclear distances, inverse internuclear distances or z-matrix coordinates are three such descriptors. The objective is to use an interpolated PES in instanton rate constant calculations, hence information on the energy, gradient, and Hessian is required at coordinates in the vicinity of the tunneling path. Instanton theory relies on smoothly fitted Hessians, therefore we use energy, gradients, and Hessians in the training procedure. A major challenge is presented in the proper back-transformation of the output gradients and Hessians from internal coordinates to Cartesian coordinates. We perform comparisons between our method, a previous approach and on-the-fly rate constant calcuations on the hydrogen abstraction from methanol and on the hydrogen addition to isocyanic acid. © 2018Wiley Periodicals, Inc.

Folding a small protein using harmonic linear discriminant analysis

Many processes of scientific importance are characterized by time scales that extend far beyond the reach of standard simulation techniques. To circumvent this impediment, a plethora of enhanced sampling methods has been developed. One important class of such methods relies on the application of a bias that is a function of a set of collective variables specially designed for the problem under consideration. The design of good collective variables can be challenging and thereby constitutes the main bottle neck in the application of these methods. To address this problem, recently we have introduced Harmonic Linear Discriminant Analysis, a method to systematically construct collective variables as linear combinations of a set of descriptors. The method uses input information that can be gathered in short unbiased molecular dynamics simulations in which the system is trapped in the metastable states. Here, to scale up our examination of the method's efficiency, we applied it to the folding of chignolin in water. Interestingly, already before any biased simulations were run, the constructed one-dimensional collective variable revealed much of the physics that underlies the folding process. In addition, using it in metadynamics, we were able to run simulations in which the system goes from the folded state to the unfolded one and back, where to get fully converged results, we combined metadynamics with parallel tempering. Finally, we examined how the collective variable performs when different sets of descriptors are used in its construction.

Mathematical structural descriptors and mutagenicity assessment: a study with congeneric and diverse datasets$

ABSTRACTQuantitative bioactivity and toxicity assessment of chemical compounds plays a central role in drug discovery as it saves a substantial amount of resources. To this end, high-performance computing has enabled researchers and practitioners to leverage hundreds, or even thousands, of computed molecular descriptors for the activity prediction of candidate compounds. In this paper, we evaluate the utility of two large groups of chemical descriptors by such predictive modelling, as well as chemical structure discovery, through empirical analysis. We use a suite of commercially available and in-house software to calculate molecular descriptors for two sets of chemical mutagens – a homogeneous set of 95 amines, and a diverse set of 508 chemicals. Using calculated descriptors, we model the mutagenic activity of these compounds using a number of methods from the statistics and machine-learning literature, and use robust principal component analysis to investigate the low-dimensional subspaces that characterize these chemicals. Our results suggest that combining different sets of descriptors is likely to result in a better predictive model – but that depends on the compounds being modelled and the modelling technique being used.

Quantitative structure–activity relationships study of potent pyridinone scaffold derivatives as HIV-1 integrase inhibitors with therapeutic applications

Human immunodeficiency virus-1 (HIV-1) integrase appears to be a crucial target for developing new anti-HIV-1 therapeutic agents. Different quantitative structure–activity relationships (QSARs) algorithms have been used in order to develop efficient model(s) to predict the activity of new pyridinone derivatives against HIV-1 integrase. Multiple linear regression (MLR) and combined principal component analysis (PCA) with MLR have been applied to build QSAR models for a set of new pyridinone derivatives as potent anti-HIV-1 therapeutic agents. Four different approaches based on MLR method including; concrete-MLR, stepwise-MLR, concrete PCA–MLR and stepwise PCA–MLR were utilized for this aim. Twenty two different sets of descriptors containing 1613 descriptors were constructed for each optimized molecule. Comparison between predictability of the “concrete” and “stepwise” procedure in two different algorithms of MLR and PCA models indicated the advantage of the stepwise procedure over that of the simple concrete method. Although the PCA was employed for dimension reduction, using stepwise PCA–MLR model showed that the method has higher ability to predict the compounds’ activity. The stepwise PCA–MLR model showed highly validated statistical results both in fitting and prediction processes ([Formula: see text] and [Formula: see text]). Therefore, using stepwise PCA approach is suitable to remove ineffective descriptors, which results in remaining efficient descriptors for building good predictability stepwise PCA–MLR. The stepwise hybrid approach of PCA–MLR may be useful in derivation of highly predictive and interpretable QSAR models.

Themes in Judges' Sentencing Remarks for Male and Female Domestic Murderers

The aim of this study was to analyse judges' sentencing remarks in cases of domestic murder. In undertaking a qualitative analysis, the methodology of grounded theory was used, and the data emanating from the judges' sentencing remarks identified several major themes as discussed in this article. Broadly speaking, the data reflect that women are viewed more harshly than men and receive higher sentences. In addition, judges are distancing male offenders from their responsibility for their violent conduct, rendering them potential candidates for more lenient treatment in terms of sentence. The analysis also reveals that judges frequently rely on stereotypes and traditional notions of marriage, family and femininity in determining an offender's sentence. These assumptions are embedded in the different set of descriptors used by judges to describe male and female offenders.

Theoretical modeling of HPV: QSAR and novodesign with fragment approach.

Structure-activity relationships in a data set of HPV6-E1 helicase ATPase inhibitors were investigated based on two different sets of descriptors. Statistically significant four parameter Quantitative Structure-Activity Relationships (QSAR) models were constructed and validated in both cases (R(2)=0.849; R(2) cv=0.811; F=52.20; s(2)=0.25; N=42). A Fragment based QSAR (FQSAR) approach was applied for developing a fragment-QSAR equation, which enabled the construction of virtual structures for novel ATPase inhibitors with desired or pre-defined activity.

Dimensionality of community structure: phylogenetic, morphological and functional perspectives along biodiversity and environmental gradients

Biodiversity is a complex and multidimensional concept that characterizes variation of life on Earth. Nonetheless, most studies have examined only a few, if not just one, dimension in isolation. Herein we conduct analyses that explicitly incorporate correlations among multiple dimensions of biodiversity by characterizing morphological, phylogenetic, and functional structure of bat communities from Atlantic Forest of South America and examine degree of redundancy among these sets of descriptors. Second, we examine dimensionality (i.e. number of orthogonal dimensions) of community structure by quantitatively determining if these different sets of descriptors correspond to unique dimensions. We assess if dimensionality measured from empirical communities differs from that based on communities randomly assembled from a regional species pool. Finally, we examine whether different indices of community structure respond differently to environmental gradients spanning Atlantic Forest. We find that Atlantic Forest bat communities are highly variable in terms of morphological, phylogenetic, and functional structure. Different sets of community structure indices exhibited substantive correlations. Accordingly, dimensionality was lower than the set of six different descriptors or even the three different biological dimensions represented. Nonetheless, observed dimensionality was greater than that expected from a null model of assembly. Only abundance‐based indices of phylogenetic structure exhibited significant environmental gradients. Temperature seasonality was the strongest predictor of phylogenetic structure, with overdispersed communities characterizing more seasonal environments and underdispersed communities occurring in areas of lower variation in temperature. Dimensionality of community structure is low with phylogenetic structure exhibiting the strongest patterns, probably because phylogeny reflects many different ecological aspects of the phenotype that are not restricted to just one index of structure. Temperature seasonality is an important determinant of phylogenetic structure of bat communities in Atlantic Forest. This research helps us to better understand the factors underlying the distribution of biodiversity, which is increasingly important for endangered ecoregions such as Atlantic Forest.

A Novel Strategy of Structural Similarity Based Consensus Modeling.

A novel strategy of "structural similarity based consensus modeling" (SSCM) based on "model distance and guided model selection" (MD-QGMS) submodel set was proposed. The SSCM strategy is built upon a hypothesis, that is, similar compounds are most probably predicted more accurately by a same submodel among a model population, which can be concluded from the fact that models employing a different set of descriptors can predict compounds with specific structures more accurately. It is proved that the proposed SSCM strategy can remarkably improve the external prediction ability of QSAR models by employing two different datasets. In future, the proposed SSCM strategy may provide a new direction to develop more accurate predictive models.

How survey design affects self-assessed health responses in the Survey of Health, Ageing, and Retirement in Europe (SHARE)

This paper considers the role of question order and framing in evaluating subjective health assessment responses using the Survey of Health, Ageing and Retirement in Europe (SHARE) dataset. In the first wave of this dataset respondents were twice asked to evaluate their health on a five-point scale, using two different sets of descriptors to define the five points, with the ordering of which set was first administered determined randomly. We document differences in inference in comparing those that were asked one question first versus those that were asked the other. We then consider determinants of the degree of concordance in responses to the two questions, as well as the characteristics of individuals that provide conflicting responses. Consistent with previous research for England (Bowling and Windsor, 2008. Journal of Epidemiology and Community Health 62, 81–85), there is evidence for eleven countries in Europe that individuals’ assessments of their health in response to the second question may be influenced by the battery of health questions that were asked following the first assessment.

Mode of action prediction of ligands of steroid hormone receptors

The several predictive models based on two well-known methods PASS and SIMCA were created. These models predict a type of physiological response of steroid compounds binding to nuclear receptors of steroid hormones. We considered 10 variants: the agonists and the antagonists of estrogen, progesterone, androgen, glucocorticoid and mineralocorticoid receptors respectively. Two different sets of descriptors were used during SIMCA (the Dragon descriptors and indices of similarity). The results of discriminant analysis are good enough with average accuracy of 80-85%.

Development of validated QSPR models for impact sensitivity of nitroaliphatic compounds

The European regulation of chemicals named REACH implies the assessment of a large number of substances based on their hazardous properties. However, the complete characterization of physico-chemical, toxicological and eco-toxicological properties by experimental means is incompatible with the imposed calendar of REACH. Hence, there is a real need in evaluating the capabilities of alternative methods such as quantitative structure–property relationship (QSPR) models, notably for physico-chemical properties.In the present work, the molecular structures of 50 itroaliphatic compounds were correlated with their impact sensitivities (h50%) using such predictive models. More than 400 olecular descriptors (constitutional, topological, geometrical, quantum chemical) were calculated and linear and multi-linear regressions were performed to find accurate quantitative relationships with experimental impact sensitivities. Considering different sets of descriptors, four predictive models were obtained and two of them were selected for their predictive reliability. To our knowledge, these QSPR models for the impact sensitivity of nitroaliphatic compounds are the first ones being rigorously validated (both internally and externally) with defined applicability domains. They hence follow all OECD principles for regulatory acceptability of QSPRs, allowing possible application in REACH.

QSAR models to predict mutagenicity of acrylates, methacrylates and [formula omitted], [formula omitted]-unsaturated carbonyl compounds

Objective The purpose of this study is to develop a quantitative structure–activity relationship (QSAR) model that can distinguish mutagenic from non-mutagenic species with α , β -unsaturated carbonyl moiety using two endpoints for this activity – Ames test and mammalian cell gene mutation test – and also to gather information about the molecular features that most contribute to eliminate the mutagenic effects of these chemicals. Methods Two data sets were used for modeling the two mutagenicity endpoints: (1) Ames test and (2) mammalian cells mutagenesis. The first one comprised 220 molecules, while the second one 48 substances, ranging from acrylates, methacrylates to α , β -unsaturated carbonyl compounds. The QSAR models were developed by applying linear discriminant analysis (LDA) along with different sets of descriptors computed using the DRAGON software. Results For both endpoints, there was a concordance of 89% in the prediction and 97% confidentiality by combining the three models for the Ames test mutagenicity. We have also identified several structural alerts to assist the design of new monomers. Significance These individual models and especially their combination are attractive from the point of view of molecular modeling and could be used for the prediction and design of new monomers that do not pose a human health risk.