The definition of nodes for graph theory analyses in biological systems is a nontrivial problem (48). In certain fields, the choice of node is fairly straightforward (choosing cities as nodes in studies of airline traffic, choosing individual people as nodes in studies of social networks, etc.). In other cases, node designation is not clear. In neuroscience research, for example, the question of what constitutes a node is open for debate [see (49, 50) for a discussion]. Furthermore, the ability to define nodes rests on the assumptions that it is possible to do so in the system under study, it is biologically plausible, and it is useful scientifically (48). The definition of edges is also nontrivial—similar issues affecting node designation also affect edge designation [e.g., (4850)]. These issues are not merely theoretical, because there are numerous examples showing that network-level properties can change depending on how nodes and edges are defined (5153).

Hence, it rests with the researcher to define these elements of the graph in a way that is driven by the data at hand yet is constrained by known biological principles. These designations then allow the analysis of network-level properties. In time, better sampling techniques will afford higher-resolution datasets, and definitions of nodes and edges will likely change as a consequence. Nevertheless, network-level analyses of these “coarse data” still provide a tractable way to lay a foundation for future connectomic work in the lung.