Abstract
A common network model of interactions is a graph. It is a metric space in which the distance between two nodes is defined as the number of edges of the shortest path connecting the two nodes. This definition of distance assumed in many mathematical models for applications in various fields may not reflect the latent geometry that underlies the structure of the network. The most frequently assumed model of metric space is that of Euclidean space, often without verifying that this assumption is plausible or supported by experimental data. It therefore becomes very important to develop methods that are able to identify the latent geometry of a network before constructing a mathematical model of its static and dynamic properties. With this motivation, I present here a method for the determination of the latent geometry of a network and show its application on a real-life case study concerning the protein-protein network of the SARS-CoV-2 virus.