Huanxiang Liu

Diagnosing breast cancer based on support vector machines.

The Support Vector Machine (SVM) classification algorithm, recently developed from the machine learning community, was used to diagnose breast cancer. At the same time, the SVM was compared to several machine learning techniques currently used in this field. The classification task involves predicting the state of diseases, using data obtained from the UCI machine learning repository. SVM outperformed k-means cluster and two artificial neural networks on the whole. It can be concluded that nine samples could be mislabeled from the comparison of several machine learning techniques.

Prediction of the Isoelectric Point of an Amino Acid Based on GA-PLS and SVMs

The support vector machine (SVM), as a novel type of a learning machine, for the first time, was used to develop a QSPR model that relates the structures of 35 amino acids to their isoelectric point. Molecular descriptors calculated from the structure alone were used to represent molecular structures. The seven descriptors selected using GA-PLS, which is a sophisticated hybrid approach that combines GA as a powerful optimization method with PLS as a robust statistical method for variable selection, were used as inputs of RBFNNs and SVM to predict the isoelectric point of an amino acid. The optimal QSPR model developed was based on support vector machines, which showed the following results: the root-mean-square error of 0.2383 and the prediction correlation coefficient R=0.9702 were obtained for the whole data set. Satisfactory results indicated that the GA-PLS approach is a very effective method for variable selection, and the support vector machine is a very promising tool for the nonlinear approximation.

QSAR Models for the Prediction of Binding Affinities to Human Serum Albumin Using the Heuristic Method and a Support Vector Machine

The binding affinities to human serum albumin for 94 diverse drugs and drug-like compounds were modeled with the descriptors calculated from the molecular structure alone using a quantitative structure-activity relationship (QSAR) technique. The heuristic method (HM) and support vector machine (SVM) were utilized to construct the linear and nonlinear prediction models, leading to a good correlation coefficient (R2) of 0.86 and 0.94 and root-mean-square errors (rms) of 0.212 and 0.134 albumin drug binding affinity units, respectively. Furthermore, the models were evaluated by a 10 compound external test set, yielding R2 of 0.71 and 0.89 and rms error of 0.430 and 0.222. The specific information described by the heuristic linear model could give some insights into the factors that are likely to govern the binding affinity of the compounds and be used as an aid to the drug design process; however, the prediction results of the nonlinear SVM model seem to be better than that of the HM.

QSAR study of ethyl 2-[(3-methyl-2,5-dioxo(3-pyrrolinyl))amino]-4-(trifluoromethyl) pyrimidine-5-carboxylate: an inhibitor of AP-1 and NF-kappa B mediated gene expression based on support vector machines.

The support vector machine, as a novel type of learning machine, for the first time, was used to develop a QSAR model of 57 analogues of ethyl 2-[(3-methyl-2,5-dioxo(3-pyrrolinyl))amino]-4-(trifluoromethyl)pyrimidine-5-carboxylate (EPC), an inhibitor of AP-1 and NF-kappa B mediated gene expression, based on calculated quantum chemical parameters. The quantum chemical parameters involved in the model are Kier and Hall index (order3) (KHI3), Information content (order 0) (IC0), YZ Shadow (YZS) and Max partial charge for an N atom (MaxPCN), Min partial charge for an N atom (MinPCN). The mean relative error of the training set, the validation set, and the testing set is 1.35%, 1.52%, and 2.23%, respectively, and the maximum relative error is less than 5.00%.

Molecular dynamics simulation, free energy calculation and structure-based 3D-QSAR studies of B-RAF kinase inhibitors.

(V600E)B-RAF kinase is the most frequent onco-genic protein kinase mutation in melanoma and is a promising target to treat malignant melanoma. In this work, a molecular modeling study combining QM-polarized ligand docking, molecular dynamics, free energy calculation, and three-dimensional quantitative structure-activity relationships (3D-QSAR) was performed on a series of pyridoimidazolone compounds as the inhibitors of (V600E)B-RAF kinase to understand the binding mode between the inhibitors and (V600E)B-RAF kinase and the structural requirement for the inhibiting activity. 3D-QSAR models, including CoMFA and CoMSIA, were developed from the conformations obtained by QM-polarized ligand docking strategy. The obtained models have a good predictive ability in both internal and external validation. Furthermore, molecular dynamics simulation and free energy calculations were employed to determine the detailed binding process and to compare the binding mode of the inhibitors with different activities. The binding free energies calculated by MM/PBSA gave a good correlation with the experimental biological activity. The decomposition of free energies by MM/GBSA indicates the van der Waals interaction is the major driving force for the interaction between the inhibitors and (V600E)B-RAF kinase. The hydrogen bond interactions between the inhibitors with Glu501 and Asp594 of the (V600E)B-RAF kinase help to stabilize the DFG-out conformation. The results from this study can provide some insights into the development of novel potent (V600E)B-RAF kinase inhibitors.

Molecular modeling study on the resistance mechanism of HCV NS3/4A serine protease mutants R155K, A156V and D168A to TMC435

Hepatitis C virus (HCV) NS3/4A protease represents an attractive drug target for antiviral therapy. However, drug resistance often occurs, making many protease inhibitors ineffective and allowing viral replication to occur. Herein, based on the recently determined structure of NS3/4A-TMC435 complex, atomic-level models of the key residue mutated (R155K, A156V and D168A) NS3/4A-TMC435 complexes were constructed. Subsequently, by using molecular dynamics simulations, binding free energy calculation and substrate envelope analysis, the structural and energetic changes responsible for drug resistance were investigated. The values of the calculated binding free energy follow consistently the order of the experimental activities. More importantly, the computational results demonstrate that R155K and D168A mutations break the intermolecular salt bridges network at the extended S2 subsite and affect the TMC435 binding, while A156V mutation leads to a significant steric clash with TMC435 and further disrupts the two canonical substrate-like intermolecular hydrogen bond interactions (TMC435(N1-H46)⋯Arg155(O) and Ala157(N-H)⋯TMC435(O2)). In addition, by structural analysis, all the three key residue mutations occur outside the substrate envelope and selectively weaken TMC435s binding affinity without effect on its natural substrate peptide (4B5A). These findings could provide some insights into the resistance mechanism of NS3/4A protease mutants to TMC435 and would be critical for the development of novel inhibitors that are less susceptible to drug resistance.

Molecular modeling study of checkpoint kinase 1 inhibitors by multiple docking strategies and prime/MM–GBSA calculation

Developing chemicals that inhibit checkpoint kinase 1 (Chk1) is a promising adjuvant therapeutic to improve the efficacy and selectivity of DNA‐targeting agents. Reliable prediction of binding‐free energy and binding affinity of Chk1 inhibitors can provide a guide for rational drug design. In this study, multiple docking strategies and Prime/Molecular Mechanics Generalized Born Surface Area (Prime/MM‐GBSA) calculation were applied to predict the binding mode and free energy for a series of benzoisoquinolinones as Chk1 inhibitors. Reliable docking results were obtained using induced‐fit docking and quantum mechanics/molecular mechanics (QM/MM) docking, which showed superior performance on both ligand binding pose and docking score accuracy to the rigid‐receptor docking. Then, the Prime/MM–GBSA method based on the docking complex was used to predict the binding‐free energy. The combined use of QM/MM docking and Prime/MM–GBSA method could give a high correlation between the predicted binding‐free energy and experimentally determined pIC50. The molecular docking combined with Prime/MM–GBSA simulation can not only be used to rapidly and accurately predict the binding‐free energy of novel Chk1 inhibitors but also provide a novel strategy for lead discovery and optimization targeting Chk1.

In silico identification of the potential drug resistance sites over 2009 influenza A (H1N1) virus neuraminidase.

The outbreak and high speed global spread of the new strain of influenza A (H1N1) virus in 2009 poses a serious threat to the general population and governments. At present, the most effective drugs for the treatment of 2009 influenza A (H1N1) virus are neuraminidase inhibitors: mainly oseltamivir and zanamivir. The use of these two inhibitors will undoubtedly increase, and therefore it is more likely that drug-resistant influenza strains will arise. The identification of the potential resistance sites for these drugs in advance and the understanding of corresponding molecular basis to cause drug resistance are no doubt very important to fight against the new resistant influenza strains. In this study, first, the complexes of neuraminidase with the substrate sialic acid and two inhibitors oseltamivir and zanamivir were obtained by fitting them to the 3D structure of 2009 influenza A (H1N1) neuraminidase obtained by homology modeling. By using these complexes as the initial structures, molecular dynamics simulation and molecular mechanics generalized Born surface area (MM-GBSA) calculations were performed to identify the residues with significant contribution to the binding of substrate and inhibitors. By analyzing the difference of interaction profiles of substrate and inhibitors, the potential drug resistance sites for two inhibitors were identified. Parts of the identified sites have been verified to confer resistance to oseltamivir and zanamivir for influenza virus of the past flu epidemic. The identified potential resistance sites in this study will be useful for the development of new effective drugs against the drug resistance and avoid the situation of having no effective drugs to treat new mutant influenza strains.

Molecular Dynamics Simulation and Free Energy Calculation Studies of the Binding Mechanism of Allosteric Inhibitors with p38α MAP Kinase

p38 MAP kinase is a promising target for anti-inflammatory treatment. The classical kinase inhibitors imatinib and sorafenib as well as BI-1 and BIRB-796 were reported to bind in the DFG-out form of human p38α, known as type II or allosteric kinase inhibitors. Although DFG-out conformation has attracted great interest in the design of type II kinase inhibitors, the structural requirements for binding and mechanism of stabilization of DFG-out conformation remain unclear. As allosteric inhibition is important to the selectivity of kinase inhibitor, herein the binding modes of imatinib, sorafenib, BI-1 and BIRB-796 to p38α were investigated by molecular dynamics simulation. Binding free energies were calculated by molecular mechanics/Poisson-Boltzmann surface area method. The predicted binding affinities can give a good explanation of the activity difference of the studied inhibitors. Furthermore, binding free energies decomposition analysis and further structural analysis indicate that the dominating effect of van der Waals interaction drives the binding process, and key residues, such as Lys53, Gly71, Leu75, Ile84, Thr106, Met109, Leu167, Asp168, and Phe169, play important roles by forming hydrogen bond, salt bridge, and hydrophobic interactions with the DFG-out conformation of p38α. Finally, we also conducted a detailed analysis of BI-1, imatinib, and sorafenib binding to p38α in comparison with BIRB-796 exploited for gaining potency as well as selectivity of p38 inhibitors. These results are expected to be useful for future rational design of novel type II p38 inhibitors.

Molecular Basis of the Interaction for an Essential Subunit PA−PB1 in Influenza Virus RNA Polymerase: Insights from Molecular Dynamics Simulation and Free Energy Calculation

The emergence of the extremely aggressive influenza recently has highlighted the urgent need for new effective treatments. The influenza RNA-dependent RNA polymerase (RdRp) heterotrimer including PA, PB1 and PB2 has crucial roles in viral RNA replication and transcription. The highly conserved PB1 binding site on PA can be considered as a novel potential drug target site. The interaction between PB1 binding site and PA is crucial to many functions of the virus. In this study, to understand the detailed interaction profile and to characterize the binding hot spots in the interactions of the PA-PB1 complex, an 8 ns molecular dynamics simulation of the subunit PA-PB1 combined with MM-PBSA (molecular mechanics Poisson-Boltzmann surface area), MM-GBSA (molecular mechanics generalized Born surface area) computations and virtual alanine scanning were performed. The results from the free energy decomposition indicate that the intermolecular van der Waals interaction and the nonpolar solvation term provide the driving force for binding process. Through the pair interaction analysis and virtual alanine scanning, we identified the binding hot spots of PA and the basic binding motif of PB1. This information can provide some insights for the structure-based RNA-dependent RNA polymerase inhibitors design. The identified binding motif can be used as the starting point for the rational design of small molecules or peptide mimics. This study will also lead to new opportunities toward the development of new generation therapeutic agents exhibiting specificity and low resistance to influenza virus.