A classification study of respiratory Syncytial Virus (RSV) inhibitors by variable selection with random forest

Microbiological and epidemiological features of respiratory syncytial virus

The properties of the main surface proteins and the viral cycle of the respiratory syncytial virus (RSV) make it an attractive pathogen from the perspective of microbiology. The virus gets its name from the manner it infects cells, which enables it to produce syncytia, which allow the virus' genetic material to move across cells without having to release viral offspring to the cellular exterior, reducing immune system identification. This causes a disease with a high impact in both children and adults over 60, which has sparked the development of several preventive interventions based on vaccines and monoclonal antibodies for both age groups. The epidemiological characteristics of this virus, which circulates in epidemics throughout the coldest months of the year and exhibits a marked genetic and antigenic drift due to its high mutation capability, must be taken into consideration while using these preventive methods. The most important microbiological and epidemiological elements of RSV are covered in this study, along with how they have affected the creation of preventive medications and their use in the future.

Rhizoma Atractylodis Macrocephalae-Assessing the influence of herbal processing methods and improved effects on functional dyspepsia

Background: The unique pharmaceutical methods for the processing of botanical drugs according to the theory of traditional Chinese medicine (TCM) affect clinical syndrome differentiation and treatment. The objective of this study was to comprehensively elucidate the principles and mechanisms of an herbal processing method by investigating the alterations in the metabolites of Rhizoma Atractylodis Macrocephalae (AMR) processed by Aurantii Fructus Immaturus (AFI) decoction and to determine how these changes enhance the efficacy of aqueous extracts in treating functional dyspepsia (FD). Methods: A qualitative analysis of AMR before and after processing was conducted using UPLC-Q-TOF-MS/MS, and HPLC was employed for quantitative analysis. A predictive analysis was then conducted using a network analysis strategy to establish a botanical drug-metabolite-target-disease (BMTD) network and a protein-protein interaction (PPI) network, and the predictions were validated using an FD rat model. Results: A total of 127 metabolites were identified in the processed AMR (PAMR), and substantial changes were observed in 8 metabolites of PAMR after processing, as revealed by the quantitative analysis. The enhanced aqueous extracts of processed AMR (PAMR) demonstrate improved efficacy in treating FD, which indicates that this processing method enhances the anti-inflammatory properties and promotes gastric motility by modulating DRD2, SCF, and c-kit. However, this enhancement comes at the cost of attenuating the regulation of motilin (MTL), gastrin (GAS), acetylcholine (Ach), and acetylcholinesterase (AchE). Conclusion: Through this series of investigations, we aimed to unravel the factors influencing the efficacy of this herbal formulation in improving FD in clinical settings.

Prediction of skin sensitization potency using machine learning approaches

The replacement of animal use in testing for regulatory classification of skin sensitizers is a priority for US federal agencies that use data from such testing. Machine learning models that classify substances as sensitizers or non-sensitizers without using animal data have been developed and evaluated. Because some regulatory agencies require that sensitizers be further classified into potency categories, we developed statistical models to predict skin sensitization potency for murine local lymph node assay (LLNA) and human outcomes. Input variables for our models included six physicochemical properties and data from three non-animal test methods: direct peptide reactivity assay; human cell line activation test; and KeratinoSens™ assay. Models were built to predict three potency categories using four machine learning approaches and were validated using external test sets and leave-one-out cross-validation. A one-tiered strategy modeled all three categories of response together while a two-tiered strategy modeled sensitizer/non-sensitizer responses and then classified the sensitizers as strong or weak sensitizers. The two-tiered model using the support vector machine with all assay and physicochemical data inputs provided the best performance, yielding accuracy of 88% for prediction of LLNA outcomes (120 substances) and 81% for prediction of human test outcomes (87 substances). The best one-tiered model predicted LLNA outcomes with 78% accuracy and human outcomes with 75% accuracy. By comparison, the LLNA predicts human potency categories with 69% accuracy (60 of 87 substances correctly categorized). These results suggest that computational models using non-animal methods may provide valuable information for assessing skin sensitization potency. Copyright © 2017 John Wiley & Sons, Ltd.

Multivariate models for prediction of human skin sensitization hazard

One of the Interagency Coordinating Committee on the Validation of Alternative Method's (ICCVAM) top priorities is the development and evaluation of non-animal approaches to identify potential skin sensitizers. The complexity of biological events necessary to produce skin sensitization suggests that no single alternative method will replace the currently accepted animal tests. ICCVAM is evaluating an integrated approach to testing and assessment based on the adverse outcome pathway for skin sensitization that uses machine learning approaches to predict human skin sensitization hazard. We combined data from three in chemico or in vitro assays - the direct peptide reactivity assay (DPRA), human cell line activation test (h-CLAT) and KeratinoSens™ assay - six physicochemical properties and an in silico read-across prediction of skin sensitization hazard into 12 variable groups. The variable groups were evaluated using two machine learning approaches, logistic regression and support vector machine, to predict human skin sensitization hazard. Models were trained on 72 substances and tested on an external set of 24 substances. The six models (three logistic regression and three support vector machine) with the highest accuracy (92%) used: (1) DPRA, h-CLAT and read-across; (2) DPRA, h-CLAT, read-across and KeratinoSens; or (3) DPRA, h-CLAT, read-across, KeratinoSens and log P. The models performed better at predicting human skin sensitization hazard than the murine local lymph node assay (accuracy 88%), any of the alternative methods alone (accuracy 63-79%) or test batteries combining data from the individual methods (accuracy 75%). These results suggest that computational methods are promising tools to identify effectively the potential human skin sensitizers without animal testing. Published 2016. This article has been contributed to by US Government employees and their work is in the public domain in the USA.

Revealing the binding mode between respiratory syncytial virus fusion protein and benzimidazole-based inhibitors

Human respiratory syncytial virus (HRSV) is a major respiratory pathogen in newborn infants and young children and can also be a threat to some elderly and high-risk adults with chronic pulmonary disease and the severely immunocompromised. The RSV fusion (RSVF) protein has been an attractive target for vaccine and drug development. Experimental results indicate a series of benzimidazole-based inhibitors which target RSVF protein to inhibit the viral entry of RSV. To reveal the binding mode between these inhibitors and RSVF protein, molecular docking and molecular dynamics simulations were used to investigate the interactions between the inhibitors and the core domain of RSVF protein. MD results suggest that the active molecules have stronger π-π stacking, cation-π, and other interactions than less active inhibitors. The binding free energy between the active inhibitor and RSVF protein is also found to be significantly lower than that of the less active one using MM/GBSA. Then, Comparative Molecular Field Analysis (CoMFA) and Comparative Molecular Similarity Indices Analysis (CoMSIA) methods were used to construct three dimensional quantitative structure-activity (3D-QSAR) models. The cross-validated q(2) values are found to be 0.821 and 0.795 for CoMFA and CoMSIA, respectively. And the non-cross-validated r(2) values are 0.973 and 0.961. Ninety-two test set compounds validated these models. The results suggest that these models are robust with good prediction abilities. Furthermore, these models reveal possible methods to improve the bioactivity of inhibitors.

Toward the prediction of FBPase inhibitory activity using chemoinformatic methods

Currently, Chemoinformatic methods are used to perform the prediction for FBPase inhibitory activity. A genetic algorithm-random forest coupled method (GA-RF) was proposed to predict fructose 1,6-bisphosphatase (FBPase) inhibitors to treat type 2 diabetes mellitus using the Mold² molecular descriptors. A data set of 126 oxazole and thiazole analogs was used to derive the GA-RF model, yielding the significant non-cross-validated correlation coefficient r²_ncv and cross-validated r²_cv values of 0.96 and 0.67 for the training set, respectively. The statistically significant model was validated by a test set of 64 compounds, producing the prediction correlation coefficient r²_pred of 0.90. More importantly, the building GA-RF model also passed through various criteria suggested by Tropsha and Roy with r²_o and r²_m values of 0.90 and 0.83, respectively. In order to compare with the GA-RF model, a pure RF model developed based on the full descriptors was performed as well for the same data set. The resulting GA-RF model with significantly internal and external prediction capacities is beneficial to the prediction of potential oxazole and thiazole series of FBPase inhibitors prior to chemical synthesis in drug discovery programs.

A systematic prediction of multiple drug-target interactions from chemical, genomic, and pharmacological data

In silico prediction of drug-target interactions from heterogeneous biological data can advance our system-level search for drug molecules and therapeutic targets, which efforts have not yet reached full fruition. In this work, we report a systematic approach that efficiently integrates the chemical, genomic, and pharmacological information for drug targeting and discovery on a large scale, based on two powerful methods of Random Forest (RF) and Support Vector Machine (SVM). The performance of the derived models was evaluated and verified with internally five-fold cross-validation and four external independent validations. The optimal models show impressive performance of prediction for drug-target interactions, with a concordance of 82.83%, a sensitivity of 81.33%, and a specificity of 93.62%, respectively. The consistence of the performances of the RF and SVM models demonstrates the reliability and robustness of the obtained models. In addition, the validated models were employed to systematically predict known/unknown drugs and targets involving the enzymes, ion channels, GPCRs, and nuclear receptors, which can be further mapped to functional ontologies such as target-disease associations and target-target interaction networks. This approach is expected to help fill the existing gap between chemical genomics and network pharmacology and thus accelerate the drug discovery processes.

Classification of HCV NS5B polymerase inhibitors using support vector machine

Using a support vector machine (SVM), three classification models were built to predict whether a compound is an active or weakly active inhibitor based on a dataset of 386 hepatitis C virus (HCV) NS5B polymerase NNIs (non-nucleoside analogue inhibitors) fitting into the pocket of the NNI III binding site. For each molecule, global descriptors, 2D and 3D property autocorrelation descriptors were calculated from the program ADRIANA.Code. Three models were developed with the combination of different types of descriptors. Model 2 based on 16 global and 2D autocorrelation descriptors gave the highest prediction accuracy of 88.24% and MCC (Matthews correlation coefficient) of 0.789 on test set. Model 1 based on 13 global descriptors showed the highest prediction accuracy of 86.25% and MCC of 0.732 on external test set (including 80 compounds). Some molecular properties such as molecular shape descriptors (InertiaZ, InertiaX and Span), number of rotatable bonds (NRotBond), water solubility (LogS), and hydrogen bonding related descriptors performed important roles in the interactions between the ligand and NS5B polymerase.