Integrating Synthetic Biology Derived Variables into Ecological Risk Assessment Using the Bayesian Network - Relative Risk…

Ecological risk assessment has not been conducted for the proposed environmental applications of synthetic biology. To develop a quantitative framework for risk assessment of synthetic biology, I selected Southeast Farallon Island as a case study for modeling the deployment of gene drive modified house mice to reduce impacts to threatened species. Southeast Farallon Island is part of the Farallon Islands National Wildlife Refuge. The island is populated by invasive house mice that impact indigenous species. Gene drive technology has been proposed as a method to suppress invasive rodent populations through CRISPR-mediated genome editing. I applied the Bayesian Network - Relative Risk Model to predict the outcomes of a gene drive mouse eradication on Southeast Farallon Island. The Bayesian Network - Relative Risk Model is able to probabilistically evaluate multiple causal pathways, incorporating the influence of multiple stressors on multiple endpoints. I used a modified version of the R-based model MGDrivE to simulate population genetics and population dynamics of gene drive mouse eradication strategies. I conducted simulations for three unique gene drive mouse release schemes across two assumed gene drive homing rates and two levels of supplemental rodenticide use; for a total of twelve simulated scenarios. I compared the relative eradication probabilities of these scenarios within a Bayesian network. Sensitivity analyses were conducted to compare the relative influence of rodenticide use and homing rate on the probability of successful mouse eradication. I found that increasing the assumed homing rate of the gene drive had a higher influence on mouse eradication than the addition of supplemental rodenticide. For most scenarios, simulations showed successful mouse eradications as early as seven years after gene drive deployment, with high probabilities of eradication within ten years.