PDF

Description

I. This thesis describes the design, functionalities, and underlying mathematics of GeneDrive.jl, a software library developed to study the effect of biotic and abiotic interactions on biological systems and optimize their control. Named in honor of a new technological horizon in geneticbased biocontrol, the tool is nonetheless broadly applicable to investigations of stage-structured metapopulation dynamics.

II. The open-source GeneDrive.jl package furnishes a three-part framework for the exploration and analysis of scenarios wherein organisms are subjected to anthropogenic and environmental change. Its components include (i) a data model to store information unique to genetic and ecological details, (ii) a dynamic model comprised of ordinary differential equations, and (iii) a decision model that discretizes the system of ODEs to formulate a nonlinear mathematical program. Written in the Julia programming language, GeneDrive.jl enables replicable, scalable, and extensible computational experiments by drawing on several state-of-the-art tools within the Julia ecosystem, including its package manager, capacity for multiple dispatch, and the robust suite of numerical solution methods available in OrdinaryDiffEq.jl and JuMP.jl respectively.

III. A principal novelty of GeneDrive.jl is the facility and conceptual clarity it offers in moving scientific experimentation from the wet lab to the dry lab to the field: the data model permits a one-time specification of empirical data on which both ODE and optimization solving algorithms can be called, allowing for the iterative testing of constraints and objectives in the decision model followed by the deployment of optimal policies in the dynamic model. GeneDrive.jl is the first biocontrol-relevant software to feature such a “testbed” functionality. It is also the first with the capacity for optimization, opening this domain area to the opportunities afforded by operations research methods.

IV. An example workflow demonstrates data model, decision model, and dynamic model creation and application. It shows the replacement of vector populations with genetically modified mosquitos that cannot carry disease, followed by an exhibition of the adjustments necessary when the same intervention is conducted under environmental perturbation. This workflow showcases how the inherent composability of GeneDrive.jl can be deployed for relevance to a range of ecological applications.

Details

Files

Statistics

from
to
Export
Download Full History