Wellcome SHD: mathematical modelling and data analysis of animal disease

Project overview

This project stems from the remainder of a Wellcome-funded Henry Dale fellowship and involves a multidisciplinary team working in three key areas: mathematical modelling, global animal disease modelling and sero-survey data analysis.

Our aims

For the mathematical modelling component, we will:

• Develop and parameterize models of rodent populations and infection dynamics using detailed rodent capture-mark-recapture data and climate data. In the first instance, we will use Lassa Fever as a case study.
• Validate these models against human case data to assess their suitability for capturing seasonal peaks in human rodent-borne disease outbreaks.
• Identify the key ecological components of rodent population dynamics that are generalisable across a range of rodent reservoir host species.
• Assess the limit of this generalisability for forecasting the timing of outbreaks of various rodent-borne diseases.

For the Global Animal Disease Modelling, we will:

• Develop workflows and tools to process species distribution information automatically and select environmental information for species distribution modelling (SDM).
• Create SDM to evaluate the likely distribution and overlap of important species for the spread of animal-borne disease, their overlap, and their potential impact on human communities.
• Evaluate these models against independent data to validate their predictive power and forecast the distribution of species important for human health under different scenarios of environmental change.
• Identify hot spots of diversity and change and their impact on human health and disease control.

Who is involved?

The Mathematical Modelling component is led by postdoctoral researcher Gregory Milne. This area focuses on the population and infection dynamics of rodents, with an initial focus on Lassa fever.

Global Animal Disease Modelling, led by Gonzalo Robles, aims to develop a global dataset and portal to inform policy on animal diseases.

The Sero-Survey Data Analysis is handled by research assistants, Ana Martinez-Checa Guiote and Harry Gordon, who are analysing data to explore risk drivers in rodent populations.

Our methods

Mathematical modelling

The mathematical modelling component seeks to understand the general ecological principles that govern the population and infection dynamics of rodents. This workstream is guided by the idea that although ecologically diverse, most rodents exhibit “fast” life histories that allow them to respond rapidly to changes in food availability.

These resource “pulses” can be driven by changes in climate. For instance, rainfall leads to sudden vegetative growth in normally arid regions. They can also be driven by rodent populations themselves. For example, falling population densities will reduce competition and increase relative resource availability.

By developing climate and demography-driven models, this workstream will identify how past changes in climate act to influence present changes in rodent populations.

To achieve this, models will be calibrated against detailed long-term data from rodent trapping studies to estimate how climate and other ecologically relevant factors influence changes in rodent population and infection dynamics.

Modelling estimates of seasonal peaks in rodent infections will then be compared to human case data to assess the utility of estimates for forecasting the timing of outbreaks of rodent-borne diseases.

The project will begin by assessing the suitability of this modelling framework for capturing seasonal outbreaks of Lassa Fever in Nigeria. Subsequently, we will aim to assess how generalisable the framework is for capturing the dynamics of a diverse range of other important rodent reservoir host species.

Global Animal Disease Modelling

With Global Animal Disease Modelling, we want to understand how the distribution of host and vector species of animal-borne diseases is related to spillover probabilities and disease prevalence.

To reconstruct the geographic distribution of our focus species, we will use the principles of “species ecological niches”, which state that the niche of a species is constant in time (not considering evolutionary time scales) but mutable in space.

This way, using a set of known locations for our species and spatially explicit environmental factors, we can model and project the likely niche and distribution of any species.

To do this, we will use species spatial records from the Global Biodiversity Information Facility (Gbif), and spatial environmental data recorded from satellite and other remote sensing techniques.

This information will then be combined into our workstream and used to calculate the distribution of important host and vector species. It will also be used to estimate the areas of overlap or likely host and vector species for specific animal-borne diseases, locate hot spots of vector-host overlap, and assess human contact/interaction.

Additionally, since climate and land-use change are likely to change global biodiversity distributions and abundances, we are also interested in how these changes are going to impact host vector distributions and their subsequent impacts on disease control and human wellbeing.

The project will begin by studying the host-vector overlap of species related to the spread of Evola, using independent and historical data to validate our results. The refined workflow will aid us in expanding our analysis to a wider scale, potentially allowing us to explore the implications of ecological impacts on animal-borne disease risks.

Our role

The Museum’s role in the project will include:

• Developing models using open-source programming software.
• Extracting open-access data from the scientific literature to calibrate models.
• Validating models using data from human rodent-borne disease outbreaks, e.g. Lassa Fever data from the Nigeria Centre for Disease Control and Prevention.
• Further model development and generalisability testing for other rodent-borne diseases.

Timeframe

• Make model code freely available for others to use
• Develop open-access apps to visualise key model outputs
• Disseminate results in scientific papers and at conferences
• Use a calibrated model to forecast outbreaks of zoonotic diseases