Comparing two classes of biological distribution systems using network analysis

Abstract

Distribution networks—from vasculature to urban transportation pathways—are spatially embedded networks that must route resources efficiently in the face of pressures induced by the costs of building and maintaining network infrastructure. Such requirements are thought to constrain the topological and spatial organization of these systems, but at the same time, different kinds of distribution networks may exhibit variable architectural features within those general constraints. In this study, we use methods from network science to compare and contrast two classes of biological transport networks: mycelial fungi and vasculature from the surface of rodent brains. These systems differ in terms of their growth and transport mechanisms, as well as the environments in which they typically exist. Though both types of networks have been studied independently, the goal of this study is to quantify similarities and differences in their network designs. We begin by characterizing the structural backbone of these systems with a collection of measures that assess various kinds of network organization across topological and spatial scales, ranging from measures of loop density, to those that quantify connected pathways between different network regions, and hierarchical organization. Most importantly, we next carry out a network analysis that directly considers the spatial embedding and properties especially relevant to the function of distribution systems. We find that although both the vasculature and mycelia are highly constrained planar networks, there are clear distinctions in how they balance tradeoffs in network measures of wiring length, efficiency, and robustness. While the vasculature appears well organized for low cost, but relatively high efficiency, the mycelia tend to form more expensive but in turn more robust networks. As a whole, this work demonstrates the utility of network-based methods to identify both common features and variations in the network structure of different classes of biological transport systems.

Substrate and cell fusion influence on slime mold network dynamics

“…We computed the shortest path between every pair of nodes along the network (using Matlab’s implementation of Dijkstra’s algorithm) in terms of edge length, and its lower bound, the Euclidean distance between node coordinates. Then, we calculated the length efficiency index (LE) defined 38 as follows: where and are the path length and the Euclidean distance between nodes i and j , respectively.…”

“…In addition, we studied the network resiliency in terms of the fault tolerance (FT) of the graphs 38 – 40 . The FT corresponds to the percentage of edges that can be removed from the graph while still connecting a given percentage of the nodes, usually .…”

How to pare a pair: Topology control and pruning in intertwined complex networks

“…Unfortunately, few metrics provide the means to fit or estimate the applied parameters of adaptation models for real systems, even though time-lapse experiments [7], counting pruning events, and topology analysis on pruned structures [3,37,38] allow for qualitative insights into the mechanism at hand for certain model organisms. Yet there has been no proposal to our knowledge to quantitatively acquire or fit the model parameters from real, pruned network structures.…”

Optimizing biologically inspired transport networks by control

“…In natural and especially biological systems, networks that transport energy, chemicals, nutrients, or materials are ubiquitous [19][20][21]. Examples include leaf veins of plants that deliver and spread nutrients, vascular systems of animals that carry oxygen to the whole body through blood circulation, and river networks that govern the flow of water in certain geographical regions [7,[22][23][24]. Such "nature-designed" networks in general function well in terms of minimizing the "cost" or dissipation and maximizing transport efficiency with reasonable fault tolerance [25].…”