Network-based rational prediction of drug combinations
For this network-based approach for drug combinations to be effective, we need to establish if the topological relationship between two drug–target modules reflects biological and pharmacological relationships, while also quantifying their network-based relationship between drug targets and HCoV-associated host proteins (drug–drug–HCoV combinations). To identify potential drug combinations, we combined the top lists of drugs. Then, “separation” measure SAB was calculated for each pair of drugs A and B using the following method:3 \documentclass[12pt]{minimal} 				\usepackage{amsmath} 				\usepackage{wasysym} 				\usepackage{amsfonts} 				\usepackage{amssymb} 				\usepackage{amsbsy} 				\usepackage{mathrsfs} 				\usepackage{upgreek} 				\setlength{\oddsidemargin}{-69pt} 				\begin{document}$$S_{AB} = \left\langle {d_{AB}} \right\rangle - \frac{{\left\langle {d_{AA}} \right\rangle + \left\langle {d_{BB}} \right\rangle }}{2},$$\end{document}SAB=dAB−dAA+dBB2,where \documentclass[12pt]{minimal} 				\usepackage{amsmath} 				\usepackage{wasysym} 				\usepackage{amsfonts} 				\usepackage{amssymb} 				\usepackage{amsbsy} 				\usepackage{mathrsfs} 				\usepackage{upgreek} 				\setlength{\oddsidemargin}{-69pt} 				\begin{document}$$\left\langle {d_ \cdot } \right\rangle$$\end{document}d⋅ was calculated based on the “closest” method. Our key methodology is that a drug combination is therapeutically effective only if it follows a specific relationship to the disease module, as captured by Complementary Exposure patterns in targets’ modules of both drugs without overlapping toxic mechanisms28.