Support vector machines in HTS data mining: Type I MetAPs inhibition study

This article reports a successful application of support vector machines (SVMs) in mining high-throughput screening (HTS) data of a type I methionine aminopeptidases (MetAPs) inhibition study. A library with 43,736 small organic molecules was used in the study, and 1355 compounds in the library with 40% or higher inhibition activity were considered as active. The data set was randomly split into a training set and a test set (3:1 ratio). The authors were able to rank compounds in the test set using their decision values predicted by SVM models that were built on the training set. They defined a novel score PT50, the percentage of the test set needed to be screened to recover 50% of the actives, to measure the performance of the models. With carefully selected parameters, SVM models increased the hit rates significantly, and 50% of the active compounds could be recovered by screening just 7% of the test set. The authors found that the size of the training set played a significant role in the performance of the models. A training set with 10,000 member compounds is likely the minimum size required to build a model with reasonable predictive power.     

In silico design of potent EGFR kinase inhibitors using combinatorial libraries

This paper is an attempt to design 4-anilinoquinazoline compounds having promising anticancer activities against epidermal growth factor (EGFR) kinase inhibition, using virtual combinatorial library approach. Partial least squares method has been applied for the development of a quantitative structure–activity relationship (QSAR) model based on training and test set approaches. The partial least squares model showed some interesting results in terms of internal and external predictability against EGFR kinase inhibition for such type of anilinoquinazoline derivatives. In virtual screening study, out of 4860 compounds in chemical library, 158 compounds were screened and finally, 10 compounds were selected as promising EGFR kinase inhibitors based on their predicted activities from the QSAR model. These derivatives were subjected to molecular docking study to investigate the mode of binding with the EGFR kinase, and the two compounds (ID 3639 and 3399) showing similar type of docking score and binding patterns with that of the existing drug molecules like erlotinib were finally reported.     

Pharmacophore modeling and virtual screening for designing potential 5-Lipoxygenase inhibitors

Inhibitors of the 5-Lipoxygenase (5-LOX) pathway have a therapeutic potential in a variety of inflammatory disorders such as asthma. In this study, chemical feature based pharmacophore models of inhibitors of 5-LOX have been developed with the aid of HipHop and HypoGen modules within Catalyst program package. The best quantitative pharmacophore model, Hypo1, which has the highest correlation coefficient (0.97), consists of two hydrogen-bond acceptors, one hydrophobic feature and one ring aromatic feature. Hypo1 was further validated by test set and cross validation method. The application of the model shows great success in predicting the activities of 65 known 5-LOX inhibitors in our test set with a correlation coefficient of 0.85 with a cross validation of 95% confidence level, proving that the model is reliable in identifying structurally diverse compounds for inhibitory activity against 5-LOX. Furthermore, Hypo1 was used as a 3D query for screening Maybridge and NCI databases within catalyst and also drug like compounds obtained from Enamine Ltd, which follow Lipinski’s rule of five. The hit compounds were subsequently subjected to filtering by docking and visualization, to identify the potential lead molecules. Finally 5 potential lead compounds, identified in the above process, were evaluated for their inhibitory activities. These studies resulted in the identification of two compounds with potent inhibition of 5-LOX activity with IC₅₀ of 14 μM and 35 μM, respectively. These studies thus validate the pharmacophore model generated and suggest the usefulness of the model in screening of various small molecule libraries and identification of potential lead compounds for 5-LOX inhibition.     

De novo design and synthesis of HIV-1 integrase inhibitors

Existing AIDS therapies are out of reach for most HIV-infected people in developing countries and, where available, they are limited by their toxicity and their cost. New anti-HIV agents are needed urgently to combat emerging viral resistance and reduce the side effects associated with currently available drugs. Toward this end, LeapFrog, a de novo drug design program was used to design novel, potent, and selective inhibitors of HIV-1 integrase. The designed compounds were synthesized and tested for in vitro inhibition of HIV-1 integrase. Out of the 25 compounds that were designed, and synthesized, four molecules (compounds 23, 26, 43, and 59) showed moderate to low inhibition of HIV-1 integrase for 3'-processing and 3'-strand transfer activities. Nonetheless, these compounds possess structural features not seen in known HIV-1 integrase inhibitors and thus can serve as excellent leads for further optimization of anti-HIV-1 integrase activity.     

Molecular modeling study on chemically diverse series of cyclooxygenase-2 selective inhibitors: generation of predictive pharmacophore model using Catalyst

Cyclooxygenase (COX) enzymes catalyse the biosynthesis of prostaglandins and thromboxane from arachidonic acid (AA). We summarize in this paper, the development of pharmacophores of a dataset of inhibitors for COX-2 by using the Catalyst/Hypogen module using six chemically diverse series of compounds. Training set consisting of 24 compounds was carefully selected. The activity spread of the training set molecules was from 0.1 to 10000 nM. The most predictive pharmacophore model (hypothesis 1), consisting of four features, namely, one hydrogen bond donor, one hydrogen bond acceptor, one hydrophobic aliphatic and one ring aromatic feature, had a correlation (r) of 0.954 and a root mean square deviation of 0.894. The entropy (configuration cost) value of the hypotheses was 16.79, within the allowed range. The difference between the null hypothesis and the fixed cost and between the null hypothesis and the total cost of the best hypothesis (hypothesis 1) was 88.37 and 78.51, respectively. The model was validated on a test set consisting of six different series of structurally diverse 22 compounds and performed well in classifying active and inactive molecules correctly. This validation approach provides confidence in the utility of the predictive pharmacophore model developed in this work as a 3D query tool in the virtual screening of drug like molecules to retrieve new chemical entities as potent COX-2 inhibitors. The model can also be used to predict the biological activities of compounds prior to their costly and time-consuming synthesis. Figure 3D Pharmacophore model generated using structurally diverse COX-2 inhibitors     

CoMFA and CoMSIA studies of angiotensin (AT1) receptor antagonists

Two 3D-QSAR methods--CoMFA and CoMSIA--were applied to a set of 38 angiotensin receptor (AT1) antagonists. The conformation and alignment of molecules were obtained by a novel method - consensus dynamics. The representation of biological activity, partial charge formalism, absolute orientation of the molecules in the grid, and grid spacing were also studied for their effect on the CoMFA models. The models were thoroughly validated through trials using scrambled activities and bootstrapping. The best CoMFA model had a cross-validated correlation coefficient ( q2) of 0.632, which improved with "region focusing" to 0.680. This model had a "predictive" r2 of 0.436 on a test series that was unique and with little representation in the training set. Although the "predictive" r2 of the best CoMSIA model, which included steric, electrostatic, and hydrogen bond acceptor fields was higher than that of the best CoMFA model, the other statistical parameters like q2, r2, F value, and s were unsatisfactory. The contour maps generated using the best CoMFA model were used to identify the structural features important for biological activity in these compounds.     

3D-QSAR and molecular docking studies of 2-pyrimidinecarbonitrile derivatives as inhibitors against falcipain-3

The three dimensional-quantitative structure activity relationship (3D-QSAR) studies were performed on a series of falcipain-3 inhibitors using comparative molecular field analysis (CoMFA) and comparative molecular similarity indices analysis (CoMSIA) techniques. A training set containing 42 molecules served to establish the QSAR models. The optimum CoMFA and CoMSIA models obtained for the training set were statistically significant with cross-validated correlation coefficients r(cv)(2) (q(2)) of 0.549 and 0.608, and conventional correlation coefficients (r(2)) of 0.976 and 0.932, respectively. An independent test set of 12 molecules validated the external predictive power of both models with predicted correlation coefficients (r(pred)(2)) for CoMFA and CoMSIA as 0.697 and 0.509, respectively. The docking of inhibitors into falcipain-3 active site using GOLD software revealed the vital interactions and binding conformation of the inhibitors. The CoMFA and CoMSIA field contour maps agree well with the structural characteristics of the binding pocket of falcipain-3 active site, which suggests that the information rendered by 3D-QSAR models and the docking interactions can provide guidelines for the development of improved falcipain-3 inhibitors.     

BCR-ABL tyrosine kinase inhibitor pharmacophore model derived from a series of phenylaminopyrimidine-based (PAP) derivatives

To reveal novel insights into the inhibition of BCR-ABL tyrosine kinase, pharmacophore mapping studies were performed for a series of phenylaminopyrimidine-based (PAP) derivatives, including imatinib (Gleevec). A seven-point pharmacophore model with one hydrophobic group (H), two hydrogen bond donors (D) and four aromatic rings (R) was developed using phase (pharmacophore alignment & scoring engine). The pharmacophore hypothesis yielded a statistically significant 3D-QSAR model, with a correlation coefficient of 0.886 and a survival score of 4.97 for training set molecules. The model showed excellent predictive power, with a correlation coefficient of Q² = 0.768 for an external test set of ten molecules. The results obtained from our studies provide a valuable tool for designing new lead molecules with potent activity.     

QSAR modeling of globulin binding affinity of corticosteroids using AM1 calculations

A quantitative structure-activity analysis of binding affinity of a series of 30 steroids for corticosteroid-binding globulin was performed using Wang-Ford charges of the non-hydrogen common atoms obtained from molecular electrostatic potential surface of AM1 optimized energy-minimized geometries of the compounds. Attempts were made to include lipophilicity (logP) and molar refractivity (MR) values of the whole molecules in the multivariate relations. The final relations were subjected to 'leave-one-out' cross-validation to check their predictive potential. It was found from the study that the charges of different atoms of the steroid nucleus [atoms 3, 4, 5 (ring A), 8, 9 (fusion points of rings B and C) and 16 (ring D)] contribute significantly to the binding affinity. This suggests the importance of these atoms/sites for the globulin binding affinity, which is also supported by previous reports on structure-activity relations of corticosteroids. Further, molar refractivity shows parabolic relation with the binding affinity, which indicates the possibility of dispersion interactions. The statistical qualities of the final equations generated in the present study (predicted variance 77-82%; explained variance 83-87%) are better than those of some of the previously reported models.     

Ferrocenyl chalcone difluoridoborates inhibit HIV-1 integrase and display low activity towards cancer and endothelial cells

We report here the discovery of a potent series of HIV-1 integrase (IN) inhibitors based on the ferrocenyl chalcone difluoridoborate structure. Ten new compounds have been synthesized and were generally found to have similar inhibitory activities against the IN 3' processing and strand transfer (ST) processes. IC(50) values were found to be in the low micromolar range, and significantly lower than those found for the non-coordinated ferrocenyl chalcones and other ferrocene molecules. The ferrocenyl chalcone difluoridoborates furthermore exhibited low cytotoxicity against cancer cells and low morphological activity against epithelial cells.     

Quantitative analysis of the relationship between structure and antioxidant activity of tripeptides

Peptides from enzymatic hydrolysates of food proteins exhibit significant antioxidant activity. Several studies have attempted to determine the factors contributing to the antioxidant activity of peptides; however, the physicochemical properties and factors essential for the antioxidant activity of peptides are still unclear. In this study, in order to clarify the factors important for peptide antioxidant activity based on the properties of component amino acids, 55 tripeptides were synthesized from 20 natural amino acids and their antioxidant activity was measured using the Trolox equivalent antioxidant capacity (TEAC) assay system. The tripeptides were divided into two data sets: a training set comprising 50 compounds and a validated set comprising five compounds. The structure‐activity relationship of the training set was then analyzed using classical quantitative structure‐activity relationship (QSAR) analysis. The study findings demonstrate that the presence of a cysteine residue at any position, an aromatic amino acid at the C‐terminus, higher hydrophobicity of the N‐terminal residue, and smaller HOMO‐LUMO energy gap of the middle residue can significantly enhance the antioxidant activity. The activities of the five validated compounds were predicted using the constructed QSAR model, and a good correlation between measured and predicted activities was observed. The information obtained from the QSAR model could be useful for effective production of antioxidant peptides from food proteins such as egg white proteins.