Evolution of bow-tie architectures in biology

TL;DR

This study uses simulations to show that bow-tie architectures naturally evolve in multi-layered systems when the evolutionary goal is rank deficient and mutations follow a product rule, explaining their prevalence in biology.

Contribution

The paper demonstrates that bow-tie structures emerge spontaneously under specific evolutionary conditions, highlighting the role of rank deficiency and product-rule mutations in their evolution.

Findings

Bow-tie architectures evolve spontaneously under certain conditions.

Product-rule mutations promote the development of bow-tie structures.

The minimal layer width equals the rank of the evolutionary goal.

Abstract

Bow-tie or hourglass structure is a common architectural feature found in biological and technological networks. A bow-tie in a multi-layered structure occurs when intermediate layers have much fewer components than the input and output layers. Examples include metabolism where a handful of building blocks mediate between multiple input nutrients and multiple output biomass components, and signaling networks where information from numerous receptor types passes through a small set of signaling pathways to regulate multiple output genes. Little is known, however, about how bow-tie architectures evolve. Here, we address the evolution of bow-tie architectures using simulations of multi-layered systems evolving to fulfill a given input-output goal. We find that bow-ties spontaneously evolve when two conditions are met: (i) the evolutionary goal is rank deficient, where the rank corresponds to the minimal number of input features on which the outputs depend, and (ii) The effects of mutations on interaction intensities between components are described by product rule - namely the mutated element is multiplied by a random number. Product-rule mutations are more biologically realistic than the commonly used sum-rule mutations that add a random number to the mutated element. These conditions robustly lead to bow-tie structures. The minimal width of the intermediate network layers (the waist or knot of the bow-tie) equals the rank of the evolutionary goal. These findings can help explain the presence of bow-ties in diverse biological systems, and can also be relevant for machine learning applications that employ multi-layered networks.