Different Combinations Of Descriptors Research Articles

An efficient and robust Phonocardiography (PCG)-based Valvular Heart Diseases (VHD) detection framework using Vision Transformer (ViT)

Background and objectives:Valvular heart diseases (VHDs) are one of the dominant causes of cardiovascular abnormalities that have been associated with high mortality rates globally. Rapid and accurate diagnosis of the early stage of VHD based on cardiac phonocardiogram (PCG) signal is critical that allows for optimum medication and reduction of mortality rate. Methods:To this end, the current study proposes novel deep learning (DL)-based high-performance VHD detection frameworks that are relatively simpler in terms of network structures, yet effective for accurately detecting multiple VHDs. We present three different frameworks considering both 1D and 2D PCG raw signals. For 1D PCG, Mel frequency cepstral coefficients (MFCC) and linear prediction cepstral coefficients (LPCC) features, whereas, for 2D PCG, various deep convolutional neural networks (D-CNNs) features are extracted. Additionally, nature/bio-inspired algorithms (NIA/BIA) including particle swarm optimization (PSO) and genetic algorithm (GA) have been utilized for automatic and efficient feature selection directly from the raw PCG signal. To further improve the performance of the classifier, vision transformer (ViT) has been implemented levering the self-attention mechanism on the time frequency representation (TFR) of 2D PCG signal. Our extensive study presents a comparative performance analysis and the scope of enhancement for the combination of different descriptors, classifiers, and feature selection algorithms. Main Results:Among all classifiers, ViT provides the best performance by achieving mean average accuracy Acc of 99.90 % and F1-score of 99.95 % outperforming current state-of-the-art VHD classification models. Conclusions:The present research provides a robust and efficient DL-based end-to-end PCG signal classification framework for designing a automated high-performance VHD diagnosis system.

Computer-Aided drug design of new 2-amino-thiophene derivatives as anti-leishmanial agents

The leishmaniasis is a neglected disease caused by a group of protozoan parasites from the genus Leishmania whose treatment is limited, obsolete, toxic, and ineffective in certain cases. These characteristics motivate researchers worldwide to plan new therapeutic alternatives for the treatment of leishmaniasis, where the use of cheminformatics tools applied to computer-assisted drug design has allowed research to make great advances in the search for new drugs candidates. In this study, a series of 2-amino-thiophene (2-AT) derivatives was screened virtually using QSAR tools, ADMET filters and prediction models, allowing direct the synthesis of compounds, which were evaluated in vitro against promastigotes and axenic amastigotes of Leishmania amazonensis. The combination of different descriptors and machine learning methods led to obtaining robust and predictive QSAR models, which was obtained from a dataset composed of 1862 compounds extracted from the ChEMBL database, with correct classification rates ranging from 0.53 (for amastigotes) to 0.91 (for promastigotes), allowing to select eleven 2-AT derivatives, which do not violate Lipinski's rules, exhibit good druglikeness, and with probability ≤70% of potential activity against the two evolutionary forms of the parasite. All compounds were properly synthesized and 8 of them were shown to be active at least against one of the evolutionary forms of the parasite with IC50 values lower than 10 μM, being more active than the reference drug meglumine antimoniate, and showing low or no citotoxicity against macrophage J774.A1 for the most part. Compounds 8CN and DCN-83, respectively, are the most active against promastigote and amastigote forms, with IC50 values of 1.20 and 0.71 μM, and selectivity indexes (SI) of 36.58 and 119.33. Structure Activity Relationship (SAR) study was carried out and allowed to identify some favorable and/or essential substitution patterns for the leishmanial activity of 2-AT derivatives. Taken together, these findings demonstrate that the use of ligand-based virtual screening proved to be quite effective and saved time, effort, and money in the selection of potential anti-leishmanial agents, and confirm, once again that 2-AT derivatives are promising hit compounds for the development of new anti-leishmanial agents.

Simple machine-learned interatomic potentials for complex alloys

Developing data-driven machine-learning interatomic potential for materials containing many elements becomes increasingly challenging due to the vast configuration space that must be sampled by the training data. We study the learning rates and achievable accuracy of machine-learning interatomic potentials for many-element alloys with different combinations of descriptors for the local atomic environments. We show that for a five-element alloy system, potentials using simple low-dimensional descriptors can reach meV/atom-accuracy with modestly sized training datasets, significantly outperforming the high-dimensional SOAP descriptor in data efficiency, accuracy, and speed. In particular, we develop a computationally fast machine-learned and tabulated Gaussian approximation potential (tabGAP) for Mo--Nb--Ta--V--W alloys with a combination of two-body, three-body, and a new simple scalar many-body density descriptor based on the embedded atom method.

Charting the Atomic C Interaction with Transition Metal Surfaces.

Carbon interaction with transition metal (TM) surfaces is a relevant topic in heterogeneous catalysis, either for its poisoning capability, for the recently attributed promoter role when incorporated in the subsurface, or for the formation of early TM carbides, which are increasingly used in catalysis. Herein, we present a high-throughput systematic study, adjoining thermodynamic plus kinetic evidence obtained by extensive density functional calculations on surface models (324 diffusion barriers located on 81 TM surfaces in total), which provides a navigation map of these interactions in a holistic fashion. Correlation between previously proposed electronic descriptors and ad/absorption energies has been tested, with the d-band center being found the most suitable one, although machine learning protocols also underscore the importance of the surface energy and the site coordination number. Descriptors have also been tested for diffusion barriers, with ad/absorption energies and the difference in energy between minima being the most appropriate ones. Furthermore, multivariable, polynomial, and random forest regressions show that both thermodynamic and kinetic data are better described when using a combination of different descriptors. Therefore, looking for a single perfect descriptor may not be the best quest, while combining different ones may be a better path to follow.

Three-Dimensional Quantitative Structure and Activity Relationship of Flavones on Their Hypochlorite Scavenging Capacity.

Flavonoids, a class of polyphenolic substances widely present in the plant realm, are considered as ideal hypochlorite scavengers. However, to our knowledge, little study has focused on the structure-activity relationship between flavonoids and hypochlorite scavenging capacity. Herein, we report for the first time the three-dimensional quantitative structure and activity relationship (3D-QSAR) combined with comparative molecular field analysis (CoMFA) and comparative molecular similarity indices analysis (CoMSIA). Four models derived from CoMFA and CoMSIA with different combinations of descriptors were built and compared; the CoMFA model, which included both steric and electrostatic fields, showed great potential (R2 = 0.989; Q2 = 0.818) in predictive quality according to both internal and external validation criteria. Additionally, the average local ionization energy (ALIE), electrostatic potential (ESP), and orbital weighted dual descriptor (OWDD) were determined to identify the key structural moiety for scavenging capacity of flavonoids against hypochlorite. The computational results indicated that hypochlorous acid (HClO) serves as an electrophile undergoing electrophilic addition to the C6 carbon, which has the highest negative charge density, which are influenced by the functional groups on the flavones. The DFT calculated mechanism revealed the catalytic role of water of mono- and di-chlorination reactions, characterized by low activation barriers, and the involvement of neutral, instead of high-energy carbocation, intermediates.

Prediction of inhibitory activities of small molecules against Pantothenate synthetase from Mycobacterium tuberculosis using Machine Learning models

The new and novel drug molecules are of prime importance against the deadly Mycobacterium tuberculosis owing to its high resistance. The discovery of new drug molecules is cost, time, and efforts intensive in chemical research. Computational approaches, such as virtual screening and Machine Learning represent an effective alternate to predict the active compounds with appreciable accuracy. In this work, we used the true active and in-active drug candidates to train the machine learned models against one of the potent drug targets from Mycobacterium tuberculosis i.e. Pantothenate synthetase (PS). We computed 1444 descriptors from the studied molecules. Initially, twenty descriptors were shortlisted based on their significant Pearson's correlation with the -logIC50 values. Different combinations of descriptors were used to optimize the number of descriptors. Further to that different Machine Learned models were applied to develop a trained model of active molecules with a reasonable accuracy. The best performed model in terms of prediction of the activity data is proposed as a model of choice to perform the data screening experiments. The current study will help to potentiate the drug discovery process against Mycobacterium tuberculosis (Mtb).

Machine-learning Interpretation of the Correlation between Infrared Emission Features of Interstellar Polycyclic Aromatic Hydrocarbons

Abstract Supervised machine-learning models are trained with various molecular descriptors to predict infrared (IR) emission spectra of interstellar polycyclic aromatic hydrocarbons. We demonstrate that a feature importance analysis based on the random forest algorithm can be utilized to explore the physical correlation between emission features. Astronomical correlations between IR bands are analyzed as examples of demonstration by finding the common molecular fragments responsible for different bands, which improves the current understanding of the long-observed correlations. We propose a way to quantify the band correlation by measuring the similarity of the feature importance arrays of different bands, by which a correlation map is obtained for emissions in the out-of-plane bending region. Moreover, a comparison between the predictions using different combinations of descriptors underscores the strong prediction power of the extended-connectivity molecular fingerprint, and shows that the combinations of multiple descriptors of other types in general lead to improved predictivity.

Molecular Modeling of Anti HIV Drugs of Protease Inhibitor Group by CoMFA

An anti-HIV agent may exert its activity by inhibiting a variety of steps in the life cycle of virus. In recent advances medicinal chemists have focused their attention mainly on virial binding to target cells, virus-cell fusion, virus uncoating, reverse transcription of genomic RNA, viral integration, gene expression, cleavage events through protease inhibitors and virion mutarotation. In present study cleavage events through protease inhibitors have been done by CoMFA. There are a number of physical, structural and chemical descriptors which may be helpful in computer aided drug designing. Out of them hydrophobicity in particular can be easily quantified for complete molecule or even for individual substituent. 26 derivatives of cyclic urea and 8 derivatives of Isosteres belonging to non-peptidic and peptidic PR inhibitors were taken as study material and a number of MLR equations using different combinations of descriptors have been developed. On the basis of these regression values, theoretical activity of the various models is predicted which provides an insight into structural and binding features of the proposed models.

QSAR Study of Some 1,3-Oxazolylphosphonium Derivatives as New Potent Anti-Candida Agents and Their Toxicity Evaluation.

The incidence of invasive fungal infections caused by Candida spp. has increased continuously in recent decades, especially in populations of immunocompromised patients or individuals hospitalized with serious underlying diseases. Therefore, the goal of our study was the search for new potent Candida albicans inhibitors via the development of QSAR models that could speed up this search process. A number of the most promising 1,3-oxazol-4-yltriphenylphosphonium derivatives with predicted activities were synthesized and experimentally tested. Furthermore, the toxicity of the studied compounds was determined in vitro using acetylcholinesterase enzyme as a biological marker. The classification QSAR models were created using Random Forests (WEKA-RF), k-Nearest Neighbors and Associative Neural Networks methods and different combinations of descriptors on the Online Chemical Modeling Environment (OCHEM) platform. Аntifungal properties of the investigated compounds were performed using standard disk diffusion method. The enzyme inhibitory action of the compounds was determined by modified Ellman's method using acetylcholinesterase from the electric organ of Electrophorus electricus. Three classification QSAR models were developed by the WEKA-RF, k-NN and ASNN methods using the ALogPS, E-State indices and Dragon v.7 descriptors. The predictive ability of the models was tested through cross-validation, giving a balanced accuracy BA = 80-91%. All compounds demonstrated good antifungal properties against Candida spp. and slight inhibition of the acetylcholinesterase activity. The high percentage of coincidence between the QSAR predictions and the experimental results confirmed the high predictive power of the developed QSAR models that can be applied as tools for finding new potential inhibitors against Candida spp. Furthermore, 1,3-oxazol-4- yl(triphenyl)phosphonium salts could be considered as promising candidates for the treatment of candidiasis and the disinfection of medical equipment.

Ensemble of deep learning, visual and acoustic features for music genre classification

ABSTRACTIn this work, we present an ensemble for automated music genre classification that fuses acoustic and visual (both handcrafted and nonhandcrafted) features extracted from audio files. These features are evaluated, compared and fused in a final ensemble shown to produce better classification accuracy than other state-of-the-art approaches on the Latin Music Database, ISMIR 2004, and the GTZAN genre collection. To the best of our knowledge, this paper reports the largest test comparing the combination of different descriptors (including a wavelet convolutional scattering network, which has been tested here for the first time as an input for texture descriptors) and different matrix representations. Superior performance is obtained without ad hoc parameter optimisation; that is to say, the same ensemble of classifiers and parameter settings are used on all tested data-sets. To demonstrate generalisability, our approach is also assessed on the tasks of bird species recognition using vocalisation and whale detection data-sets. All MATLAB source code is available.

Predicting DPP-IV inhibitors with machine learning approaches.

Dipeptidyl peptidase IV (DPP-IV) is a promising Type 2 diabetes mellitus (T2DM) drug target. DPP-IV inhibitors prolong the action of glucagon-like peptide-1 (GLP-1) and gastric inhibitory peptide (GIP), improve glucose homeostasis without weight gain, edema, and hypoglycemia. However, the marketed DPP-IV inhibitors have adverse effects such as nasopharyngitis, headache, nausea, hypersensitivity, skin reactions and pancreatitis. Therefore, it is still expected for novel DPP-IV inhibitors with minimal adverse effects. The scaffolds of existing DPP-IV inhibitors are structurally diversified. This makes it difficult to build virtual screening models based upon the known DPP-IV inhibitor libraries using conventional QSAR approaches. In this paper, we report a new strategy to predict DPP-IV inhibitors with machine learning approaches involving naïve Bayesian (NB) and recursive partitioning (RP) methods. We built 247 machine learning models based on 1307 known DPP-IV inhibitors with optimized molecular properties and topological fingerprints as descriptors. The overall predictive accuracies of the optimized models were greater than 80%. An external test set, composed of 65 recently reported compounds, was employed to validate the optimized models. The results demonstrated that both NB and RP models have a good predictive ability based on different combinations of descriptors. Twenty "good" and twenty "bad" structural fragments for DPP-IV inhibitors can also be derived from these models for inspiring the new DPP-IV inhibitor scaffold design.

Computer Assisted QSAR/QSPR Approaches – A Review

Background/Objectives: Quantitative Structure-Activity Relationship (QSAR) / Quantitative Structure - Property Relationship (QSPR) model is based on changes in molecular structure that would reflect changes in observed biological activity or physico-chemical property. Methods/Statistical analysis: QSAR/QSPR involves chemistry, biology and statistics fields for analysis. It has been widely accepted model for predicting association between molecular structure and its activity. Over the years many algorithms have been proposed and applied in QSAR/QSPR studies. Framework of model involves molecular structure (graph) representation, calculation of molecular descriptors (graph invariants) and multiple linear regression method is applied for analysis. Model has been validated through statistical parameters (R and R2). Findings: Methods involved in model development were reviewed for QSAR/QSPR studies. Multiple Linear Regression is one of the best methods for developing QSAR/QSPR model. This work focuses on developing QSPR model for predicting boiling point of alkyl benzene molecules using Multiple Linear Regression method. Wiener index, Harary Index, Hyper Wiener Index, Hyper Harary Index, and Randic index are calculated for analysis. The model has been validated by calculating R and R2 value. Various models were developed based on different combinations of descriptors to analysis which contribute best in predicting boiling point. Best fit model has been identified by developing model with different combinations of descriptors and rank them based on highest R and R2 value. Model with highest value has been taken for prediction of boiling point as best fit model as n (number of molecules) =14, R= 0.9934 and R2=0.9968. Applications/Improvements: Review on methods involved in prediction analysis has enlightened that model with reduced molecular descriptor subset and outlier detection method shows better performance by improving quality of the dataset Main application of QSAR/QSPR analysis is in drug discovery process. As it has reduced the time taken for lead identification and optimization in drug discovery process.

Adolescentes em conflito com a lei e família: um estudo de revisão sistemática da literatura

This study conducted a systematic literature review to ascertain the state of the art on the theme “Family and delinquent adolescents” in the period 2007-2012. National and international scientific journals abstracts indexed in Bireme databases (Index Psi. Revistas Tecnico-Cientificas, PePsic, SciELO, LILACS, Medline) were selected through different combinations of descriptors. A total of 162 studies were located using the following criteria: it is within the period of analysis, it is written in English, Spanish and Portuguese and without duplicity. From the group, only 35 articles mentioned the family theme, and they were analyzed in full. There was a predominance of international studies with an empirical and a quantitative character. From the analysis three themes were identified: characteristics and functioning of delinquent adolescents family dynamics; adolescents’ perception about their families, and methodologies for assessment and intervention aimed at the family. The results revealed that it is necessary to consider the practice of assisting the off ending adolescent with the same emphasis given to his/her family. The results also revealed that it has been a common practice to blame the family for the child’s off ense.Key words: delinquent adolescents, family, systematic review.

SMALL WORLD NETWORK STRATEGIES FOR STUDYING PROTEIN STRUCTURES AND BINDING

Small world network concepts provide many new opportunities to investigate the complex three dimensional structures of protein molecules. This mini-review explores the published literature on using small-world network approaches to study protein structure, with emphasis on the different combinations of descriptors that have been tested, on studies involving ligand binding in protein-ligand complexes, and on protein-protein complexes. The benefits and success of small world network approaches, which change the focus from specific interactions to the local environment, even to non-local phenomenon, are described. The purpose is to show the different ways that small world network concepts have been used for building new computational models for studying protein structure and function, and for extending and improving existing modelling approaches.

Entheseal Changes: Do They Reflect Socioeconomic Status in the Early Medieval Central European Population? (Mikulčice – Klášteřisko, Great Moravian Empire, 9th – 10th century)

ABSTRACTEntheseal changes (EC) of skeletons are supposed to reflect the physical activity of our ancestors. Similarly, the elements of the burial rite (e.g. grave goods, burial depth) are presumed to be related to the social status of individuals. The aim of this study was to verify the existence of relationships between EC distribution and archaeological descriptors providing information on social status.Fibrocartilaginous entheses of upper and lower limb bones of 115 individuals (aged 20–50 years) from the early medieval burial site Mikulčice – Klášteřisko were evaluated using Villotte's method (2006). Factor analysis (FA) was applied two times with two different combinations of descriptors: (i) FA‐1 based on the archaeological characteristics; (ii) FA‐2 based on the distribution of EC for males and females separately. The second step was important for the validation of results acquired from archaeological data.It is possible to link several models of EC distribution with selected archaeological characteristics. (i) The relationship between EC and archaeological descriptors is more visible in males than females. (ii) Males withwarrior equipmentand deep graves very probably belonged to the higher social class and are characterised by the loading of the triceps brachii on the olecranon and of the gluteus medius on the femur. (iii) Males with shallower graves and objects of daily use more typically demonstrate an occurrence of EC at the insertion site of the wrist flexors or extensors. The clear links demonstrated in males were absent in females. There exists only a weak relationship between the descriptorjewelleryand the incidence of EC at the insertion sites of the shoulder and certain insertion sites of the lower limbs (especially the hip rotators). This poorly significant relationship supports our previous hypothesis that females from Mikulčice performed the same activities regardless their socioeconomic status. Copyright © 2012 John Wiley & Sons, Ltd.

Predictive QSAR modeling of phosphodiesterase 4 inhibitors

A series of diverse organic compounds, phosphodiesterase type 4 (PDE-4) inhibitors, have been modeled using a QSAR-based approach. 48 QSAR models were compared by following the same procedure with different combinations of descriptors and machine learning methods. QSAR methodologies used random forests and associative neural networks. The predictive ability of the models was tested through leave-one-out cross-validation, giving a Q 2 = 0.66–0.78 for regression models and total accuracies Ac = 0.85–0.91 for classification models. Predictions for the external evaluation sets obtained accuracies in the range of 0.82–0.88 (for active/inactive classifications) and Q 2 = 0.62–0.76 for regressions. The method showed itself to be a potential tool for estimation of IC 50 of new drug-like candidates at early stages of drug development.

A comparison of different QSAR approaches to modeling CYP450 1A2 inhibition

Prediction of CYP450 inhibition activity of small molecules poses an important task due to high risk of drug-drug interactions. CYP1A2 is an important member of CYP450 superfamily and accounts for 15% of total CYP450 presence in human liver. This article compares 80 in-silico QSAR models that were created by following the same procedure with different combinations of descriptors and machine learning methods. The training and test sets consist of 3745 and 3741 inhibitors and noninhibitors from PubChem BioAssay database. A heterogeneous external test set of 160 inhibitors was collected from literature. The studied descriptor sets involve E-state, Dragon and ISIDA SMF descriptors. Machine learning methods involve Associative Neural Networks (ASNN), K Nearest Neighbors (kNN), Random Tree (RT), C4.5 Tree (J48), and Support Vector Machines (SVM). The influence of descriptor selection on model accuracy was studied. The benefits of "bagging" modeling approach were shown. Applicability domain approach was successfully applied in this study and ways of increasing model accuracy through use of applicability domain measures were demonstrated as well as fragment-based model interpretation was performed. The most accurate models in this study achieved values of 83% and 68% correctly classified instances on the internal and external test sets, respectively. The applicability domain approach allowed increasing the prediction accuracy to 90% for 78% of the internal and 17% of the external test sets, respectively. The most accurate models are available online at http://ochem.eu/models/Q5747 .

Exploring the binding of HIV-1 integrase inhibitors by comparative residue interaction analysis (CoRIA)

Since the recognition of HIV-1 integrase as a novel and rational target for HIV therapeutics, remarkable progress has been made in the development of integrase inhibitors. Computational techniques have played a critical role in accelerating research in this area. However, most previous computational studies were based solely on ligand information. In the present work, we describe the application of one of our recently developed receptor-based 3D-quantitative structure activity relationships (QSAR) methods, i.e. comparative residue interaction analysis (CoRIA), in exploring the events involved in ligand-integrase binding. In this methodology, the non-bonded interaction energies (van der Waals and Coulombic) of the inhibitors with individual active site residues of the integrase enzyme are calculated and, along with other thermodynamic descriptors, are correlated with biological activity using chemometric methods. Different combinations of descriptors were used to develop three types of QSAR models, all of which were found to be statistically significant by internal and external validation. This is the first report of such a dedicated receptor-based 3D-QSAR approach being applied to comprehend the integrase-inhibitor recognition process. In addition, the study was performed on 13-different series of inhibitors, thereby exploring the most structurally diverse data set ever used in understanding the inhibition of HIV-1 integrase. The major advantage of this technique is that it can quantitatively extract crucial residues and identify the nature of interactions between the ligand and receptor that modulate activity. The models suggest that Asp64, Thr66, Val77, Asp116, Glu152 and Lys159 are the key residues influencing the binding of ligands with the integrase enzyme, and the majority of these results are in line with earlier studies. The approach facilitates easy lead-to-hit conversion and design of novel inhibitors by optimisation of the interaction of ligands with these specific residues of the integrase enzyme.