Abstracts 2024/25
Abstracts Term 2
20 January 2025: Valaire Yatat (University of Yaounde 1, Cameroon)
Title: Modelling the Impact of the Sterile insect technique with accidental releases of sterile females on mosquito-borne diseases control when viruses are circulating
Abstract: The sterile insect technique (SIT) is a technique to control some vectors of diseases by releasing sterile males. However, during these releases, sterilized females can be (accidentally) released and since only females are vectors of diseases, it is important to study their impact when arthropod viruses are circulating. To that aim, we develop and study an entomological-epidemiological model, considering either permanent or periodic releases.
Our results reveal that outside an epidemic period, the release of sterile females is not an issue, as long as the sterile males release rate is greater than $\Lambda_M^{crit}$}. Within an epidemic period, we show that sterile females releases do not really impact the SIT efficiency, as long as the release rate, $\Lambda_F$, is lower than a critical value, $\Lambda_F^{crit}$, that depends on the mosquito and epidemiological threshold parameters, $\N$, and $\R_0^2$. To illustrate numerically our theoretical results, we consider Dengue parameters. We estimate all thresholds and also the effective reproduction number, $\R_{eff}^2$, and highlight the importance of early permanent or periodic SIT control to prevent or mitigate the risk of a Dengue epidemic, with and without sterile female releases.
27 January 2025: Edward Offord & Hakan Ferhatosmanoglu (University of Warwick)
Edward Offord
Title: Single cell time series analysis using graph neural networks
Abstract: The analysis of individual biological cells is important to understand how many processes in the human body function. Software is required to parse the large volumes of data collected. Graph neural networks offer a generalized method for extracting useful information from time-series data. As an example, Dictyostelium preforming 'cell drinking' (shared by many human cancer cells) is used to demonstrate the potential of these networks to aid in our understanding of subcellular processes. 'Cell drinking' is characterized by an invagination that develops on the surface of the cell; here the graph neural network detects and extracts a complete life cycle of the surface deformation. The invaginations (called macropinocytic cups) then provide insight into how the surface deformation is formed over time. Understanding the mechanism in greater detail enables researchers to experiment with the development of drugs that can either excite or inhibit the formations of these structures.
Hakan Ferhatosmanoglu
Title: Scalable Graph Neural Networks: Systems and Applications
Abstract: Graph Neural Networks (GNNs) jointly encode graph topology and node features to perform tasks such as classification, link prediction, and forecasting. While they are applicable across various domains, such as recommendation systems and scientific modelling, scaling GNNs to large datasets and deeper architectures remains challenging. This talk outlines our distributed GNN training and inference system to address these challenges, involving a (hyper)graph partitioner that uses sparsity for scalability and a stream processor to maintain up-to-date embeddings for evolving graphs. I will explain how the solution seeks to mitigate issues such as oversmoothing in GNNs and the potential neighbourhood explosion during distributed training and inference. Finally, I will present a forecasting system we built which involves a spatio-temporal GNN model on data collected from multiple sensors.
3 February 2025: Jessica Clark (University of Glasgow)
Title: Understanding mechanisms of inter-epidemic Rift Valley Fever virus transmission
Abstract: TBC
10 February 2025: Rebecca Hoyle (University of Southampton)
Title: Multidisciplinary Ecosystem to study Lifecourse Determinants and Prevention of Early-onset Burdensome Multimorbidity
Abstract: TBC
17 February 2025: Michael Head (University of Southampton)
Title: Neglected tropical diseases, climate change and entering the Chief’s Palace: health needs, challenges and anecdotes around Last Mile populations in rural Ghana
Abstract: TBC
24 February 2025: Reiko Tanaka (Imperial College London)
Title: TBC
Abstract: TBC
3 March 2025: Joaquin Prada (University of Surrey)
Title: Informing Cystic Echinococcosis control programs through Mathematical Modelling
Abstract: Mathematical and Statistical models have been used extensively to inform public health and animal health policy in the last decades. In this talk, we will go over some recent methods that have been applied in the context of Cystic Echinococcosis, a neglected zoonotic disease affecting predominantly remote farming communities. We will explore also how we might need to go beyond these models to better support policy decision-making.
10 March 2025: Luke Davis (University of Edinburgh)
Title: TBC
Abstract: TBC
Abstracts Term 1
7 October 2024: Francesca Scarabel (University of Leeds)
Title: Numerical methods for structured population models in ecology and epidemiology
Abstract: In this talk I will consider mathematical models for populations where individual rates are completely determined by a continuous structuring variable that evolves in time (e.g. age or size in ecology, age or age of infection in epidemiology). They can be described as renewal equations or partial differential equations of transport type, and the set of software tools available for these types of equations is much more limited compared to that available for compartmental models formulated as ordinary differential equations (ODEs). In recent years, within a collaboration with the University of Udine (Italy), I have developed a series of user-friendly numerical methods to study the stability and bifurcations of structured population models by means of a convenient approximation with ODEs, which can be studied with well-established software for ODEs. More recently, similar numerical techniques have been used to obtain an efficient method to approximate the reproduction numbers. I will illustrate the methods with applications to mathematical ecology and epidemiology.
14 October 2024: Emma Davis (University of Warwick)
Title: Applications of branching processes to disease emergence and elimination
21 October 2024: Denis Patterson (University of Durham)
Title: Spatial models of forest-savanna bistability
Abstract: Empirical studies suggest that for vast tracts of land in the tropics, closed-canopy forests and savannas are alternative stable states, a proposition with far-reaching implications in the context of ongoing climate change. Consequently, numerous spatially implicit and explicit mathematical models have been proposed to capture the mechanistic basis of this bistability and quantify the stability of these ecosystems. We present an analysis of a spatially extended version of the so-called Staver-Levin model of forest-savanna dynamics (a system of nonlinear partial integro-differential equations). On a homogeneous domain, we uncover various types of pattern-forming bifurcations in the presence of resource limitation, which we study as a function of the resource constraints and length scales in the problem. On larger (continental) spatial scales, heterogeneity plays a significant role in determining observed vegetative cover. Incorporating domain heterogeneity leads to interesting phenomena such as front-pinning, complex waves, and extensive multi-stability, which we investigate analytically and numerically.
28 October 2024: Anne Skeldon (University of Surrey)
Title: Mathematical modelling of the sleep-wake cycle: light, clocks and digital-twins
Abstract: We all sleep. But what determines when and for how long? In this talk I’ll describe some of the fundamental mechanisms that regulate sleep. I’ll introduce the nonsmooth coupled oscillator systems that form the basis of current mathematical models of sleep-wake regulation and discuss their dynamical behaviour. I will describe how we are using models to unravel environmental, societal and physiological factors that determine sleep timing and outline how constructing digital-twins could enable us to create personalised light interventions for sleep timing disorders.
4 November 2024: Xander O'Neill (Heriot-Watt University)
Title: Pathogen persistence in wildlife populations
Abstract: How do highly virulent pathogens persist? We start by delving into the dynamics of African swine fever, a highly virulent pathogen, which can be sustained in a wild boar population despite a mortality rate of 90-100%. How does this persist? How could the introduction of this virus impact other, more chronic illnesses, such us tuberculosis? Can the lack of control for one make it easier to control or eradicate the other? This comes full circle when we propose a more general study, asking the question, what key model characteristics lead to slower (or faster) approximate times to extinction?
11 November 2024: Laura Wadkin (Newcastle University)
Title: Modelling the spread of tree diseases and invasive pests through UK treescapes
Abstract: The loss of biodiversity due to the spread of destructive tree diseases and invasive pests within our native forests is having an enormous environmental, economic, and social impact. In the ‘25 Year Environment Plan’ the UK government highlights enhancing biosecurity as a key priority, through the control of existing diseases and pests, and by building forest resilience against new ones. We are working in collaboration with Defra to develop mathematical models to deepen our understanding of the fundamental behaviours of key pests and pathogens, act as predictive tools for forecasting, and to explore different control strategies. Broadly, we use a combination of partial differential equations, agent-based modelling, and statistical inference techniques. In this talk I will give an overview of the collaborative work to date and present a case study example of the oak processionary moth epidemic in London parks to show how the parameters for a compartmental SIR model with a time varying infection rate can be inferred.
18 November 2024: Weini Huang (Queen Mary University of London)
Title: Mathematical models of extra-chromosomal DNA and their applications in cancer
Abstract: Many diseases in human including cancers are caused by genetic alternations/errors starting from a single cell. The origin of these genetic errors and the expansion of the abnormal cells carrying these genetic errors are often stochastic processes. Here we develop a general framework to model the dynamics of cancer cells carrying extra-chromosomal DNA (ecDNA), a genetic error found in more than 30% of tumour samples across various cancer types and correlated to the worse clinical outcomes. Different from chromosomal DNA where genetic materials are on average equally divided to daughter cells controlled by centromeres during mitosis, the segregation of ecDNA copies is random partition and leads to a fast accumulation of cell-to-cell heterogeneity in copy numbers. We use deterministic and stochastic approaches to analyse the fraction of cells carrying ecDNA and copy number distributions, and use those distributions observed in experimental and clinical data to infer the selection strength of ecDNA positive cells. We further extend our model of a single ecDNA species where all ecDNA copies are identical in genetic structure and function, to multiple ecDNA types where ecDNA copies can have different genes (species), mutations (genotypes), or have different functions without genetic changes (phenotypes). All these variations of our basic model can be applied to relevant biological context and provide insight to understand ecDNA dynamics observed in clinic or experiments and improve treatment strategies.
25 November 2024: Matt Keeling (University of Warwick)
Title: Cost-effectiveness of COVID-19 vaccination
Abstract: not abstract -- Matt kindly agreed to give a talk at very short notice after a cancellation.
2 December 2024: Nardus Mollentze (University of Glasgow)
Title: Predicting spatial expansions in the risk of virus spillover from vampire bats
Abstract: Common vampire bats are distributed throughout much of Latin America, where their obligate blood feeding lifestyle creates a high risk for cross-species transmission of viruses to humans and livestock. While it remains difficult to study virus transmission directly in the bat population, thousands of spillovers of rabies virus to livestock reveals distinctive signatures of both endemic circulation and epidemic spread in the reservoir. I will discuss recent work combining Bayesian phylogeography and fine-scale mixture models of the rate and probability of invasions to examine the factors predicting spatial expansions in the areas experiencing spillovers. I will also briefly describe the ongoing development of a massively multiplexed serology assay which which will allow us to directly track the circulation of all known bat-associated viruses. By improving our understanding of the spatial spread of viruses in this key reservoir, these studies are bringing us closer to the long-awaited goal of predicting spillover risk in space and time.