A generative model for bipartite gene-sharing networks

Abstract

Significance Gene sharing among viruses leaves a characteristic imprint on the structure of bipartite gene-genome networks, yet the evolutionary mechanisms underlying their topological features are not well understood. We present a mechanistic generative model for bipartite gene-sharing networks that jointly predicts the degree distributions of genes and genomes from three fundamental processes: horizontal gene transfer, functional innovation, and organismal innovation. Analytical results and simulations show that this simple model reproduces the contrasting statistical patterns observed in viral and microbial pangenome data. By fitting the model to empirical networks, we obtain quantitative estimates of innovation rates and show that gene gain dominates over gene loss in virus evolution. This work connects elementary evolutionary events with the large-scale architecture of gene-sharing networks.