A principal component regression model for predicting phytochemical binding to the H. pylori CagA protein

In this study, a predictive model for the binding free energy of natural compounds towards the cagA protein is presented. The formulated model which is built on principal component —multiple linear regression demonstrates reliable accuracy (r2test = 0.92, RMSEtest = 0.483), while only requiring five independent variables for the prediction. It was further noted that topological descriptors had the greatest influence on the generated principal components which served as the predictors. The created regression model can help promote and accelerate the disco very of natural compounds as cagA binders for the development of anti-H. pylori agents.
Source: Network Modeling Analysis in Health Informatics and Bioinformatics - Category: Bioinformatics Source Type: research