Low dimensionality of phenotypic space as an emergent property of coordinated teams in biological regulatory networks

Abstract

Biological networks driving cell-fate decisions involve complex interactions, but they often give rise to only a few phenotypes, thus exhibiting low-dimensional dynamics. The network design principles that govern such cell-fate canalization remain unclear. Here, we investigate networks across diverse biological contexts- Epithelial-Mesenchymal Transition, Small Cell Lung Cancer, and Gonadal cell-fate determination - to reveal that the presence of two mutually antagonistic, well-coordinated teams of nodes leads to low-dimensional phenotypic space such that the first principal component (PC1) axis can capture most of the variance. Further analysis of artificial team-based networks and random counterparts of biological networks reveals that the principal component decomposition is determined by the team strength within these networks, demonstrating how the underlying network structure governs PC1 variance. The presence of low dimensionality in corresponding transcriptomic data confirms the applicability of our observations. We propose that team-based topology in biological networks are critical for generating a cell-fate canalization landscape.