3D-QSARpy: Combining variable selection strategies and machine learning techniques to build QSAR models

Abstract Quantitative Structure-Activity Relationship (QSAR) is a computer-aided technology in the field of medicinal chemistry that seeks to clarify the relationships between molecular structures and their biological activities. Such technologies allow for the acceleration of the development of new compounds by reducing the costs of drug design. This work presents 3D-QSARpy, a flexible, user-friendly and robust tool, freely available without registration, to support the generation of QSAR 3D models in an automated way. The user only needs to provide aligned molecular structures and the respective dependent variable. The current version was developed using Python with packages such as scikit-learn and includes various techniques of machine learning for regression. The diverse techniques employed by the tool is a differential compared to known methodologies, such as CoMFA and CoMSIA, because it expands the search space of possible solutions, and in this way increases the chances of obtaining relevant models. Additionally, approaches for select variables (dimension reduction) were implemented in the tool. To evaluate its potentials, experiments were carried out to compare results obtained from the proposed 3D-QSARpy tool with the results from already published works. The results demonstrated that 3D-QSARpy is extremely useful in the field due to its expressive results.

3D QSAR and docking studies on benzoylsulfonohydrazides as histone acetyltransferase KAT6A inhibitors

Sixty-one analogs of benzoylsulfonohydrazides were subjected to 3D QSAR studies using CoMFA and CoMSIAtechniques followed by docking studies to develop a correlation of the structure with their respective activities. Thegenerated model had shown good predictability and the contour analysis followed by docking study has provided aninsight to develop new inhibitors. The cross-validation values corresponding to CoMFA and COMSIA were observedto be within the acceptable criterion (q2 > 0.5). The docking analysis of the best active compound shown was −41.81kcal/mol. From the obtained analysis results of CoMFA as well as CoMSIA, the data can be useful to develop morepotent histone acetyltransferase inhibitors.

Automated QSAR and Docking Study of Phyto-Ligands of Chromolaena odorata as Inhibitor of Bacteria Beta-lactamase Enzyme

Bacteria resistance to antibacterial antibiotics is made possible by theproduction of beta-lactamase. Beta-lactamase enzyme confers resistance by breakingopen the Beta-lactam structure of antibiotics, thereby deactivating their antibacterialproperties. As a result of this, attention shifted into identifying potential lead inhibitorof beta-lactamase, with ability to reduce resistance encountered in bacteria antibiotics.The computational approach was employed in the generation of QSAR model usingAutomated QSAR, and in the docking of ligands from Chromolaena odorata with Betalactamase. The best model obtained was KPLS_Dendritic_43 (R2 = 0.8564 andQ2=0.7819), and was used in predicting the bioactivity of the lead compounds. Dockingstudy revealed that the ligands bind with a higher binding score than co-crystallizedligand and other standard drug employed in this study. Tianshic acid and chromomoraterecorded binding energy of -9.305 and -7.989 respectively. The drug-like properties ofthe ligands were evaluated using the Lipinski rule of Five, which revealed that C. odorataligands do not only inhibit the activity of beta-lactamase, but the ligands are also druglike. Therefore, further studies are needed to adequately justify the mechanism of actionof these ligands as a beta-lactamase inhibitor.

3D-QSAR, molecular docking and dynamics simulation of difluorophenol pyridine derivatives as RSK2 inhibitor

Given the increasing role of P90 Ribosomal S6 Kinase 2 (RSK2) as an anticancer drug target, we performed3D-Quantitative structure–activity relationship, including comparative molecular field analysis (CoMFA) andcomparative molecular similarity indices analysis (CoMSIA) on difluorophenol pyridine derivatives as the inhibitorof RSK2. CoMFA model with q2 of 0.597 and R2 of 0.993, while CoMSIA model with q2 of 0.563 and R2 of 0.993,were obtained. The predictive ability of both models was assured using a test set compound with R2pred values of 0.996each. Using the validated models, novel compound was proposed and its interaction with RSK2 was investigatedemploying molecular docking and molecular dynamics simulation of 50 ns. Furthermore, molecular-mechanicsPoisson–Boltzmann surface area calculation was performed. The result showed that the newly designed compoundhas a comparable binding free energy with the known RSK2 inhibitor, indicating its potential as a new RSK2 inhibitor.

3d Qsar And Docking Study Of Indole Derivatives As Selective Cox-2 Inhibitors

Objective: Non-steroidal anti-inflammatory agents (NSAIDs) continue to be one of the most widely used groups of therapeutic agents. QSAR (quantitative structure-activity relationship) approach is a very useful and widespread technique for drug design. 3D QSAR facilitates evaluation of three-dimensional molecular fields around molecules and generates a relationship of these fields' values with the activity. Methods: 3D QSAR study was performed on selected twenty-four compounds from synthesized indole derivatives using the stepwise variable selection k-nearest neighbor (kNN) molecular field analysis approach for indicating the contribution of the steric and electronic field for activity. The docking study was performed to further confirm the binding affinity of synthesized molecules (ligands) to COX-2 enzyme as well as to study binding nature. Results: Statistically significant model was generated using VLife Molecular Design Suite 3.5 software with cross-validated correlation coefficient q2 of 0.9461 and high predictive correlation coefficient (Pred_r2) of 0.8782 indicating that the model is robust. The results of docking study suggest that the synthesized compounds have a comparable binding affinity with the COX-2 enzyme. Conclusion: The present study may prove to be helpful in the development and optimization of existing indole derivatives as anti-inflammatory agents with selective COX-2 inhibition.

3D-QSAR, molecular docking, and dynamics simulation of quinazoline–phosphoramidate mustard conjugates as EGFR inhibitor

To develop novel and more potent quinazoline–phosphoramidate mustard conjugates as epidermal growth factorreceptor (EGFR) inhibitor, three-dimensional quantitative structure-activity relationship [comparative molecularfield analysis (CoMFA) and comparative molecular similarity indices analysis (CoMSIA)] combined with moleculardocking were performed. A series of 13 compounds in the training set gave q2 values of 0.577 and 0.537, as well as r2values of 0.926 and 0.921 for CoMFA and CoMSIA models, respectively. The contour maps that were produced by theCoMFA and CoMSIA models revealed that steric, electrostatic, and hydrophobic fields were crucial in the inhibitoryactivity of quinazoline–phosphoramidate derivatives. Based on the CoMFA and CoMSIA models, several novel EGFRinhibitors were designed, which established crucial interactions at the ligand binding domain of EGFR. Nearly, 100ns MD simulation indicated the stability of the designed compounds at 100 ns, while molecular mechanics-PoissonBoltzmann surface area calculation showed that the designed compound had a higher affinity than that of the parentcompound.

3D-QSAR Study for Pyridine Heterocyclic Ring PI3K Inhibitor as Anti-renal Cancer Drug / 中国药房

OBJECTIVE:To study 3D-QSAR of pyridine heterocyclic ring PI3K inhibitor as anti-renal cancer drug,and to provide reference for the design and R&D of new anti-renal cancer inhibitors. METHODS:The data of structure and active value (pIC50) of 30 pyridine heterocyclic ring PI3K inhibitors were collected. After Sybyl-X 1.1 software used for molecular superimposition, CoMFA and CoMSIA model were established to investigate three dimensional field, electrostatic field, hydrophobic field,hydrogen bond donor site and hydrogen bond acceptor field of PI3K inhibitor molecule. Sybyl-X 1.1 software was used for molecular docking,and the mechanism of PI3K inhibitor molecule and receptor target protein were analyzed. PyMOL V1.5 software was used to design new PI3K inhibitor molecules. The activity of inhibitor molecules was predicted with CoMFA and CoMSIA model. RESULTS:The cross validation coefficients of CoMFA and CoMSIA model were 0.617 and 0.601, fitting validation coefficients were 0.969 and 0.974,and external predictive correlation coefficients were 0.656 and 0.670, respectively. In CoMFA model, contributions of three dimensional field and electrostatic field were 56.2% and 43.8%respectively. In CoMSIA model,contributions of three dimensional field,electrostatic field,hydrophobic field,hydrogen bond donor site and hydrogen bond acceptor field were 41.0%,31.3%,21.1%,2.4%,4.2%. After molecular superimposition,small steric hindrance,strong positive and hydrophilic groups introduced nearby R1 group of common skeleton could help to enhance the activity of molecules. The results of molecular docking showed that PI3K inhibitor molecule formed three hydrogen bonds with the key amino acids ALA805,VAL882 and THR887 of receptor target protein,with the length of 1.84,1.99,1.99 ?. According to above information,6 new molecules were designed,among which predicted pIC50 of 2 molecules with higher activity were 3.211,3.247(CoMFA method)and 3.238,3.222(CoMSIA method). CONCLUSIONS:Established new CoMFA and CoMSIA model have good prediction ability and statistical stability. Contribution of three dimensional field is higher than that of electrostatic field,and the influence of hydrophobic field on molecular activity can not be ignored. Pyridine heterocyclic ring PI3K inhibitors have strong hydrogen bonding role with receptor target protein. 3D-QSAR can provide reference for the design,reconstruction and drug R&D of new PI3K inhibitor molecule.

3D-QSAR Analysis and Molecular Docking Study on Biaryl Aminothiazine BACE1 Inhibitor / 中国药房

OBJECTIVE:To provide theoretic basis for the design and synthesis of novel high-activity biaryl aminothiazineβ-amyloid precursor protein cleaving enzyme 1 (BACE1) inhibitor,the research and development of new AD therapy drugs. METHODS:Totally 41 molecules of biaryl aminothiazine BACE1 inhibitors were selected. By SYBYL-X 2.0 software package, CoMFA and CoMSIA method were used to construct 3D-QSAR model of derivatized compounds. Surflex-dock molecular docking was applied to analyze binding mode of the compounds with BACE1. RESULTS:The q2 value of 3D-QSAR model established by CoMFA and CoMSIA method were all higher than 0.5,indicating good predictability. The established three dimensional contour plots could manifest the effect of substituents at different sites on activity of compounds. Surflex-dock analysis showed that biaryl aminothiazine and amino acid residues as ASP80, ASP276 and TYR246 in BACE1 had a key effect on hydrogen bonds. CONCLUSIONS:3D-QSAR model established on the basis of biaryl aminothiazine derivatized compounds show good predictability,which provides guidance for the structure optimization of the compound. TYR246 may be another potential active functional residue of biaryl aminothiazine inhibitor compound molecule combined with BACE1. Through 3D-QSAR analysis and molecular docking,new biaryl aminothiazine BACE1 inhibitor can be designed and synthesized so as to research and develop new drugs for AD.

3D-QSAR studies on thiazole derivatives as potent inhibitors of dihydroorotate dehydrogenase / 药物评价研究

Objective The three-dimensional quantitative structure activity relationship (3D-QSAR) method was applied to study thiazole derivatives as potent inhibitors ofdihydroorotate dehydrogenase,which provided useful guidance for more discovery of potent inhibitors of dihydroorotate dehydrogenase.Methods Molecular field analysis (CoMFA) and comparative molecular similarity indices analysis (CoMSIA) were applied to systematicly investigate 3D-QSAR of 38 hiazole derivatives as potent inhibitors of dihydroorotate dehydrogenase.Established models of CoMFA and CoMSIA and the predictive ability of models were validated.Three dimensional map was applied to analyzing the relationship between structure and activity of thiazole derivatives.Results The coefficients of cross validation q2 and non-cross validation r2 for CoMFA model were 0.796 and 0.978,and for CoMSIA model were 0.721 and 0.976 respectively.The prediction of activity of compound was close to the actual value of the two models.Effect of compound structure on its activity could be analyzed comprehensively and intuitively by three dimensional map.Conclusion The model reveals the relationship between the structure characteristics and the inhibitory activity,and has good predictive capability and stability to lay a good foundation for further development and research.

Discovery of noveI EphB4 inhibitors based on pharmacophore modeIing and virtuaI screening techniques / 中国药科大学学报

The objective of this paper is to discover new potent inhibitors against EphB4 for cancer therapy via computer-aided drug design strategies including building 3D-QSAR models,virtual screening and molecular doc-king means.The first step is to generate pharmacophore models based on Catalyst/HypoGen algorithm.The best model,Hypo1,has the highest Correl value (0.96),the lowest RMS value (0.89),the closest total cost (101.26) to fixed cost (89.20),and the best Δcost (89.14).Subsequently,Hypo1 was subjected to test set validation and Fischer′s randomization verification and then was used as a 3D query to screen database.In order to further nar-row the number of hits,drug-likeness screening and molecular docking techniques were applied.Finally,23 novel molecules with diverse scaffolds were selected as possible candidates against EphB4 for further studies based on predicted activity analysis,docking scores,and binding modes analysis methods.

In silico Analysis of 3D-QSAR and Molecular Docking for Bcl-2 Inhibitors to Potential Anticancer Drug Development.

Apoptosis control is characterized by a delicate balance between homo and hetero dimerization of pro- and anti-apoptosis members of the protein family. Inhibiting this protein protein interaction is one viable approach to cancer therapy. Anti-apoptosis the prosurvival family members Bcl-2, Mcl-1, and Bcl-XL are current targets for anti-cancer drug design. The chemotherapy has aroused many researchers‟ interests and a great deal of current efforts has been focusing on the design and development of various anticancer drugs. Ligand-based drug designing methods approaches through pharmacophore mapping and Three Dimention- Quantitative Structure Activity Relationship (3D-QSAR) are used in drug discovery as well as molecular docking to seek potential binding sites of the Bcl-2 protein and its inhibitors interactions. Dynamically predictive 3D-QSAR model with Pearson-r value (0.74), F (62.5), Standard Deviation (0.285) of the regression and Root Mean Square Deviation RMSE (0.321), Q2(0.514) that was obtained for binding affinity of Bcl-2 protein respectively. The bioinformatics techniques were proved that the development of good potential activity drug compound to cancer. To our knowledge the results describes anti-tumour activity of HEQ-1 drug compound promising to convey anti-tumour drug development.

Estudos de QSAR-3D em derivados 5-nitro-2-tiofilidênicos com atividade frente a Staphylococcus aureus multi-resistente / 3D QSAR studies of 5-nitrothiophene derivatives with antimicrobial activity against multidrug-resistant Staphylococcus aureus

Campos moleculares extraídos de aplicativos utilizados em estudos de QSAR-3D apresentam, em geral, grande número de informações, muitas vezes irrelevantes na expressão da atividade biológica. O programa Volsurf converte as informações presentes em mapas de energia de interação tridimensionais em número reduzido de descritores bidimensionais que se caracterizam como de fácil entendimento e interpretação. Assim, foram avaliados, neste estudo, dezoito derivados 5-nitro-2-tiofilidênicos com atividade antimicrobiana frente a Staphylococcus aureus multi-resistente, correlacionando as características tridimensionais destes ligantes com a referida atividade. Para o desenho e conversão tridimensional dos ligantes foram utilizados os aplicativos Sybyl (Tripos Inc) e CORINA (Molecular Networks GmbH Computerchemie), respectivamente. Os campos de interação molecular foram calculados no programa GRID (Molecular Discovery Ltd). A aplicação do programa Volsurf (Molecular Discovery Ltd) resultou em modelo estatisticamente robusto (r² = 0,93, q² = 0,87) com 48 descritores estruturais, mostrando ser a hidrofobicidade propriedade fundamental no condicionamento da atividade antimicrobiana.

Studies in three-dimensional molecular fields generally contain a large amount of data, some of which are redundant or not relevant. The program Volsurf, a quite fast method, is able to compress the relevant information present in 3D molecular structures into a few easy bidimensional descriptors. This study correlates the antimicrobial activity of eighteen 5-nitro-2-thiophylidene derivatives against multidrug-resistant Staphylococcus aureus with three-dimensional molecular fields of these ligands. For molecular structures sketching and 3D conversion, Sybyl and CORINA programs were used, respectively. The GRID force field was applied to generate the 3D interaction energies. The Volsurf characterization results on significant statistic model with 48 descriptors (r² = 0,93, q²= 0,87), observing a significant influence of hydrophobic properties on antimicrobial activity performance.