Despite decades of intensive research into malaria vaccines, there is still little prospect for one with high efficacy at preventing disease. The most deadly malaria parasite in humans, Plasmodium falciparum, has a long duration of infection in the blood stage, and during this time it induces a strong natural antibody response. These antibodies are known to be effective at clearing infection in non-immune individuals, and their primary target is P. falciparum erythrocyte membrane protein 1 (PfEMP1). This protein would be an ideal vaccine target except for the fact that its diversity in natural populations is so extensive and poorly understood that approaches targeting this protein have generally been considered unrealistic. We revisit the possibility of a PfEMP1-based vaccine after considering whether PfEMP1 diversity can be simplified and understood from the perspectives of protein function,disease ecology and evolutionary dynamics. For this work we deeply sample PfEMP1 antigenic sequences from the asymptomatic reservoir of infection in a highly endemic community within the Bongo District of Northern Ghana. We sample asymptomatic cases because these represent the vast majority of the parasite population and our aim is to consider the overall structure and dynamics of PfEMP1 antigenic diversity at the population level. Our pilot survey includes the PfEMP1 antigenic sequences from 208 parasite positive isolates, within which we observe over 10,000 distinct PfEMP1 sequence types—the majority of which only occur once in the dataset. We decompose this diversity into its constituent parts and then functionally classify the sequences based on their composition, which allows us to reduce the number of antigenic types by almost two orders of magnitude. We also consider whether the PfEMP1 diversity is non-randomly structured into strains by natural selection in response to competitive dynamics, and we find evidence for this in the form of reduced antigenic overlap between isolates and non-random modularity among antigenic sequences not explainable by neutral forces. Finally,we use a simulation modeling approach to identify some of the unique features of antigenic evolution by recombination, which are likely relevant to PfEMP1-based interventions since this protein is highly reliant on recombination for the creation of new variants. Considering the lengthy and strong exposure of the immune system to PfEMP1, if we can achieve a better understanding of its natural diversity, we believe this protein has the potential to be a highly effective vaccine target and that PfEMP1-based vaccines should be considered alongside other malaria interventions in the effort to achieve elimination.
The mammary gland is comprised of interacting cell types presumed to have distinct gene expression patterns relating to their cellular origins, function and spatio-temporal context. Yet, its cellular composition is understood principally from studies of mixed cell populations without detailed knowledge of the differences between the individual cells comprising it. This makes it difficult to understand how the tissue maintains its homoeostasis and balance of cell types and how this is perturbed in cancer development. We are addressing this problem by generating single cell transcriptomes from fetal, post-natal and adult mouse mammary epithelia. The resulting data provides new challenges and opportunities in understanding cell identity. These include dealing with sparse data, searching for higher order information embedded in the transcriptome as a network and understanding the relationship between heterogeneous transcriptional profiles and cell type.
Statistical analysis of infectious outcomes is important for an optimal understanding of the kinds of interventions that can be used to limit the spread of infectious diseases. A principal challenge in the analysis of infectious diseases is the lack of fully observed data that are important for understanding the underlying dynamics of transmission. Multiple approaches have been developed that allow for missing data and misclassification. In this talk, we will review some of the previous work on modeling transmission with incomplete data. We will also discuss our recent developments of dynamic transmission models to estimate key parameters related to infectious diseases, particularly those related to antibiotic resistant bacteria.
There is concern about the lack of reproducibility of biomedical studies, but the research community has not taken advantage of formal Uncertainty Quantification (UQ) methods to better understand this issue. Here we show the importance of UQ in Translational Science and Precision Medicine. This presentation describes the use of UQ methods in biomedical research with applications to breast cancer classification, family history risk assessment and gene co-expression network determination. The results presented here show that UQ methods can be applied to biomedical sciences. UQ provides useful clinical and translational information, and arguable UQ should become a common tool in translational science and precision medicine because UQ methods can provide a better understanding of the underlying factors leading to the lack of reproducibility.
The success of science is due in large part to the hierarchical nature of physical theories. These effective theories model natural phenomena as if the physics at macroscopic length scales were almost independent of the underlying, shorter-length-scale details. The efficacy of these simplified models can be understood in terms of parameter identifiability. Parameters associated with microscopic degrees of freedom are usually unidentifiable as quantified by the Fisher Information Matrix. I apply an information geometric approach in which a microscopic, mechanistic model is interpreted as a manifold of predictions in data space. Model manifolds are often characterized by a hierarchy of boundaries—faces, edges, corners, hyper-corners, etc. These boundaries correspond to reduced-order models, leading to a model reduction technique known as the Manifold Boundary Approximation Method. In this way, effective models can be systematically derived from microscopic first principles for a variety of complex systems in physics, biology, and other fields.
In this work we develop a variational formalism for decomposing the effect of stochastic noise on traveling waves in neural fields. The noise is decomposed into two parts: a shift of the wavefront, and a small perturbation about the wavefront. We find that the decomposition is accurate on timescales that are exponential in the inverse of the strength of the noise. The decomposition is applied to various neural field waves in product spaces, and also waves that are induced by a stimulus.
Acute lymphoblastic leukemia (ALL) is the most common cancer in children, comprising approximately 25% of all childhood malignancies. ALL is characterized by the overproduction and accumulation of immature, abnormal white blood cells (lymphoblasts) and consecutive displacement of normal hematopoiesis. The backbone of maintenance therapy includes daily oral 6-Mercaptopurine and weekly oral Methotrexate.
This talk introduces a complete model of chemotherapy-induced leukopenia during maintenance therapy of ALL, taking the effects of two drugs in combination into consideration. Patient-specific parameters of the model are estimated. Furthermore, the ability of the proposed model to predict leukocytes over a given duration is justified by appropriate methods. All of the results here are conducted based on a real dataset collected from a local hospital.
Individuals within a population differ in attributes such as size, sex, spatial location, behaviour and age. Structured population models allow us to understand the effects of within population trait variation on demographic and population level processes. Data-driven modelling approaches allow parameters and functional relationships of a model to be statistically estimated from actual individual observations. This help our models reflect reality.
Integral projections models (IPM’s) are a data-driven structured population modelling approach, in which individuals are described by continuous traits, rather than being classified into arbitrary categories. They can be used to obtain quantitative and predictive inferences about a wide range of population types. They can describe secies with a very broad range of life histories, ranging from plants with dormant seed banks, to animals with size dependent growth, survival and reproduction.
I will demonstrate how IPM’s can be constructed in order to make full use of precious experimental and field observation. I will show current work that is being undertaken to advance IPM’s and I will outline future directions for the application of these models in cancer research.