Statistical model for IC50 determination of acetylcholinesterase enzyme for Alzheimer’s disease

Abstract

This study aimed to formulate a suitable statistical model to determine acetylcholinesterase enzyme's half-maximal inhibitory (IC₅₀) by a series of synthetic compounds. It was done with the same core structure for acetylcholinesterase inhibition for anti-Alzheimer’s disease (AD). The IC₅₀ of eighteen synthesized compounds on anticholinesterase activities was obtained and statistical methods were applied. Regression models were fitted to the dose-response curve to look for their IC₅₀. Simple linear regression is the simplest model for the dose-response curve. However, polynomial regression models or non-linear regression models fit the data more accurately. The adjusted coefficient of determination (𝑅²_𝑎𝑑𝑗) was used to determine the best model among the linear models, while the root mean square error (RMSE) is more suitable in determining the goodness of fit between linear and non-linear model. Four-parameter logistic (4-PLR) regression often fits the dose-response data closely. Based on the RMSE value, a polynomial regression fitted better than 4-PLR with the IC₅₀ of 245.52.