A generative model for bipartite gene-sharing networks

Area of Science:

• Evolutionary biology
• Virology
• Network science

Background:

• Gene-sharing networks are crucial for understanding virus and mobile genetic element evolution.
• These networks display distinct degree distributions: scale-free for genes and exponential for genomes.

Purpose of the Study:

• To propose a mechanistic model explaining observed degree distributions in gene-sharing networks.
• To identify fundamental evolutionary processes driving these patterns.

Main Methods:

• Developed a mechanistic model incorporating horizontal gene transfer, gene capture, genome emergence, and gene loss.
• Employed a mean-field approximation to derive analytical expressions for degree distributions.
• Validated predictions using numerical simulations and empirical data from viral and prokaryotic pangenomes.

Main Results:

• The model analytically predicts a power-law distribution for genes and an exponential distribution for genomes.
• Numerical simulations confirmed these predictions, with parameter values fitting empirical data.
• Setting gene loss to zero showed strong agreement with observed distributions, particularly for viruses.

Conclusions:

• A simple, two-parameter model effectively explains bipartite gene-sharing network structures.
• Viral evolution is predominantly driven by gene gain, aligning with independent evolutionary reconstructions.
• The model provides insights into genome plasticity and evolutionary forces shaping viral genomes.