UC San Diego

It is difficult to research the evolution of cell differentiation in nature because of the unavailability of a suitable model organism and the advanced time since the divergence of single and multicellular organisms. Instead of looking at a natural model we decided to explore this issue with a computer model. We based our model on the one proposed by Andreas Wagner in 1994 to look at the evolution of gene networks. We modified the model to account for selection goals that promoted cell differentiation. Because of the prevalence of modular components in more advanced biological systems we expected modularity to form in the evolved gene

networks. Therefore, we tested the gene networks for modularity using the K-Means clustering test. The model was completed after many editions that swapped between concepts introduced in previous scientists’ adaptations of the Wagner model. Increased modularity was observed in the evolved populations of the model over the unevolved populations, 0.149 over 0.136 p < 2.2e-16. However, this change was so minor there may be no biological significance. Additionally, whether the model evolved towards one fitness goal then the other or both at the same time was tested and found to be both. Finally, whether connectivity increased when selecting for differentiation was tested and the results are inconclusive at best.