Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Discovery Grants Program - Individual

McMaster University

$56,800.00

Semimechanistic models for ecology and epidemiology

FY 2024-25

Summary
This program will develop semimechanistic models, along with tools to encourage their adoption by a broad biological audience. Semimechanistic models embed nonparametric smooth curves such as Gaussian processes within the framework of a mechanistic model, to deepen our understanding of predator-prey and host-parasite systems. Such models should provide more interpretable, theoretically relevant information than machine learning approaches, as well as improving prediction by imposing realistic constraints that allow the models to more efficiently learn the shapes of the key functions. By comparing estimated smooth functions with standard functional forms (such as Holling-type functional responses), we can gain deeper understanding of the processes driving ecological systems. I will use semimechanistic models to characterize key nonlinear ecological interactions in two ecological systems: predator-prey interactions and epidemiological dynamics. Predator-prey interactions are driven by the functional response (FR), the per-predator rate of prey consumption as a function of prey density and (possibly) other prey traits such as size. Fitting the FR as an increasing, smooth, probably saturating function can unify analyses in this field, and could illuminate the reasons for the shape-sensitivity of predator-prey dynamics. In epidemic models, the key function is the incidence function; semimechanistic models will substitute a smooth (and smoothly time-varying) function C(S, I, t) for the usual assumption of a parametric incidence such as ßSI. The technical component of the project will implement new constrained smooth functions for use within ecological and epidemiological semimechanistic models, focusing on previously unexplored classes of constraints (such as anchoring to zero or convergence to a horizontal asymptote). Next, it will extend existing methodology to allow for inference on important properties of estimated curves (e.g. existence, location, and height of minima/maxima), a major current limitation of nonparametric estimation. It will build general software tools to allow a broad range of ecologists to build and apply semimechanistic models to their systems. The applied component of the program will assemble existing predator-prey and epidemic-curve data from the literature and use these data sets to evaluate improvements in prediction, as well as sensitivity of predictions and dynamics to functional forms. The project will develop nonparametric components that fit the requirements of ecological/epidemiological models; understand how much data of what quality is needed to make fitting semimechanistic models worthwhile; and extend and unify what we know about predator-prey and epidemic dynamics by applying these methods to the extensive body of existing data sets. We will use the lessons learned in these studies to build software that will make these methods accessible to a broad audience of quantitative population biologists.
Program
Discovery Grants Program - Individual
Location
Ontario, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data

As published

Application id
RGPIN-2023-05400
Name nom
Bolker, Benjamin
Department département
Mathematics & statistics
Organization id
27
Institution établissement
McMaster University
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2023
Award amount
56800
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1503
Committee name en
Evolution and Ecology
Committee name fr
Évolution et écologie
Area of application code
1200
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Advancement of knowledge
Area of application fr
Avancement des connaissances
Research subject code
4700
Research subject group en
Evolution and ecology
Research subject group fr
Évolution et écologie
Research subject en
Evolution and ecology
Research subject fr
Évolution et écologie
Field of research list names en
Not available
Field of research list names fr
Non disponible
Socioeconomic objective list names en
Not available
Socioeconomic objective list names fr
Non disponible
Application title
Semimechanistic models for ecology and epidemiology
Keywords
predator-prey; host-parasite; parameter estimation; statistical modeling; infectious disease; dynamics
Application summary
This program will develop semimechanistic models, along with tools to encourage their adoption by a broad biological audience. Semimechanistic models embed nonparametric smooth curves such as Gaussian processes within the framework of a mechanistic model, to deepen our understanding of predator-prey and host-parasite systems. Such models should provide more interpretable, theoretically relevant information than machine learning approaches, as well as improving prediction by imposing realistic constraints that allow the models to more efficiently learn the shapes of the key functions. By comparing estimated smooth functions with standard functional forms (such as Holling-type functional responses), we can gain deeper understanding of the processes driving ecological systems. I will use semimechanistic models to characterize key nonlinear ecological interactions in two ecological systems: predator-prey interactions and epidemiological dynamics. Predator-prey interactions are driven by the functional response (FR), the per-predator rate of prey consumption as a function of prey density and (possibly) other prey traits such as size. Fitting the FR as an increasing, smooth, probably saturating function can unify analyses in this field, and could illuminate the reasons for the shape-sensitivity of predator-prey dynamics. In epidemic models, the key function is the incidence function; semimechanistic models will substitute a smooth (and smoothly time-varying) function C(S, I, t) for the usual assumption of a parametric incidence such as ßSI. The technical component of the project will implement new constrained smooth functions for use within ecological and epidemiological semimechanistic models, focusing on previously unexplored classes of constraints (such as anchoring to zero or convergence to a horizontal asymptote). Next, it will extend existing methodology to allow for inference on important properties of estimated curves (e.g. existence, location, and height of minima/maxima), a major current limitation of nonparametric estimation. It will build general software tools to allow a broad range of ecologists to build and apply semimechanistic models to their systems. The applied component of the program will assemble existing predator-prey and epidemic-curve data from the literature and use these data sets to evaluate improvements in prediction, as well as sensitivity of predictions and dynamics to functional forms. The project will develop nonparametric components that fit the requirements of ecological/epidemiological models; understand how much data of what quality is needed to make fitting semimechanistic models worthwhile; and extend and unify what we know about predator-prey and epidemic dynamics by applying these methods to the extensive body of existing data sets. We will use the lessons learned in these studies to build software that will make these methods accessible to a broad audience of quantitative population biologists.