Date of Award

2019

Publication Type

Doctoral Thesis

Degree Name

Ph.D.

Department

Computer Science

Keywords

biological invasions, ecology, evolution, genetic diversity, individual-based model, sequence processing

Supervisor

Robin Gras

Supervisor

Hugh MacIsaac

Rights

info:eu-repo/semantics/openAccess

Abstract

Biological invasions are characterized by the movement of organisms from their native geographic region to new, distinct regions in which they may have significant impacts. Biological invasions pose one of the most serious threats to global biodiversity, and hence significant resources are invested in predicting, preventing, and managing them. Biological systems and processes are typically large, complex, and inherently difficult to study naturally because of their immense scale and complexity. Hence, computational modelling and simulation approaches can be taken to study them. In this dissertation, I applied computer simulations to address two important problems in invasion biology. First, in invasion biology, the impact of genetic diversity of introduced populations on their establishment success is unknown. We took an individual-based modelling approach to explore this, leveraging an ecosystem simulation called EcoSim to simulate biological invasions. We conducted reciprocal transplants of prey individuals across two simulated environments, over a gradient of genetic diversity. Our simulation results demonstrated that a harsh environment with low and spatially-varying resource abundance mediated a relationship between genetic diversity and short-term establishment success of introduced populations rather than the degree of difference between native and introduced ranges. We also found that reducing Allee effects by maintaining compactness, a measure of spatial density, was key to the establishment success of prey individuals in EcoSim, which were sexually reproducing. Further, we found evidence of a more complex relationship between genetic diversity and long-term establishment success, assuming multiple introductions were occurring. Low-diversity populations seemed to benefit more strongly from multiple introductions than high-diversity populations. Our results also corroborated the evolutionary imbalance hypothesis: the environment that yielded greater diversity produced better invaders and itself was less invasible. Finally, our study corroborated a mechanical explanation for the evolutionary imbalance hypothesis – the populations evolved in a more intense competitive environment produced better invaders. Secondly, an important advancement in invasion biology is the use of genetic barcoding or metabarcoding, in conjunction with next-generation sequencing, as a potential means of early detection of aquatic introduced species. Barcoding and metabarcoding invariably requires some amount of computational DNA sequence processing. Unfortunately, optimal processing parameters are not known in advance and the consequences of suboptimal parameter selection are poorly understood. We aimed to determine the optimal parameterization of a common sequence processing pipeline for both early detection of aquatic nonindigenous species and conducting species richness assessments. We then aimed to determine the performance of optimized pipelines in a simulated inoculation of sequences into community samples. We found that early detection requires relatively lenient processing parameters. Further, optimality depended on the research goal – what was optimal for early detection was suboptimal for estimating species richness and vice-versa. Finally, with optimal parameter selection, fewer than 11 target sequences were required in order to detect 90% of nonindigenous species.

Download

Share

COinS