A Reexamination of Noise-Driven Robustness Evolution in Gene Regulatory Networks
Gene regulatory network (GRN) models are widely used to investigate gene-expression dynamics in biological systems. In this study, we re-examine an evolutionary GRN model proposed by Kaneko, which was used to investigate the relationship between robustness to nongenetic phenotypic noise and robustness to mutation. We explicitly reconstruct the simulation protocol by specifying computational details omitted from the original study and reproduce its main qualitative findings, including the noise-dependent suppression of low-fitness individuals, the relationship between isogenic phenotypic variance and genetic variance, and the increased robustness of evolved networks under certain expression noise. We further analyze the robustness of evolved populations to variations in initial conditions. While the representative-network analysis shows quantitative differences from the original study, the population-level analysis supports the qualitative conclusion that evolution under higher phenotypic noise leads to broader basins of attraction. We provide a reusable implementation and publicly release the source code to support reproducibility and facilitate further studies on robustness in dynamical GRN models.