Analyzing nonlinear patterns between entropy and graph indices through regression models in Anoctamin Network

Abstract

This article gives an extensive regression-based characterization of degree-dependent topological indices and entropy quantification for the Anoctamin II (AnO2) network, a calcium-activated chloride channel plays a crucial role in neural signaling and perception processes. Utilizing logarithmic and quadratic regression approaches, we quantitatively investigate correspondence between structural graph indices and respective entropy descriptors. Our findings consistently show that logarithmic regression is superior to quadratic in describing the network structure-dependent nonlinear relationships contained in the molecular graph. From statistical modeling and data visualization, we uncover unique growth patterns among different indices including Randic, Zagreb, and Augmented Zagreb with respective entropy quantification. These findings hold significant implications for quantifying molecular complexity and modeling predictive bio-structure behavior in network-based pharmaceutical design, materials science, and computational biology applications.