Multi-scale modelling of COVID-19 transmission, with application to new and on-going vaccination strategies

我打算从事的工作将与数学科学研究领域有关，因为这项工作是通过数学方程的视角研究生物系统的。该项目的结论可能会影响未来疫苗推广的发展。其核心概念是利用数学建模来确定个体在多次感染SARS-CoV-2和接种疫苗后抗体库的短期演变。所使用的数学模型将是简洁和直观的，重点是参数的可识别性。目前的想法是使用捕食者-猎物模型结构来研究抗体库的发展，主要竞争发生在淋巴组织中的幼稚b细胞和记忆b细胞之间。该模型将由从已知感染史的个体收集的详细抗体数据提供信息。随着项目的进展，我打算考虑感染和接种疫苗的顺序，以及个体感染的变种。此外，交叉反应性抗体已经在人群中存在的假设将被检验。这些个体水平模型的结果将用于为旨在了解重复感染和疫苗接种如何保护人群免受疾病和感染的群体水平模型提供信息。这项分析的目的是获得有关SARS-CoV-2保护寿命的信息，并检验关于它是否会流行以及以何种形式流行的假设。这项研究的新颖之处在于所采用的方法及其对大量可用数据的使用。在过去对SARS-CoV-2感染的研究中，没有一个宿主内模型考虑到淋巴组织中竞争性b细胞发育与随后b细胞产生的抗体对病毒的回收之间的相互作用。此外，现有的宿主内模型的应用主要是为了了解疾病进展和住院风险，从而集中对宿主的损害。因此，现有的模型往往没有明确地模拟免疫过程，而只是模拟它们对病毒和宿主细胞的影响，或者即使它们模拟免疫过程，它们也是与细胞损伤相关的，如细胞因子和细胞毒性t细胞。非SARS-CoV-2特异性的旧模型试图捕捉b细胞发育的过程。然而，这些模型忽略了感染部位，并对淋巴组织外的动态做出了不准确的假设。此外，这些旧的模型通常不能用数据验证它们的结论。然而，我的简约模型捕捉到了淋巴组织中的竞争和感染部位的相关动态，并且是可观察的，因此可以通过数据验证。这些特性使其更有助于了解疫苗接种的效果，并为今后的疫苗接种运动提供信息。

The work I intend to undertake will be related to the mathematical sciences area of research, as the work studies biological systems through the lens of mathematical equations. Conclusions of the project could affect the development of roll out of future vaccines. The core concept is to use mathematical modelling to determine the short term evolution of individual's antibody repertoires through multiple SARS-CoV-2 infection and vaccinations. The mathematical models used will be parsimonious and intuitive, with a focus on parameter identifiability. The current idea is to use a predator-prey model structure for studying the development of antibody repertoires with the main competition being between naive and memory B-cells in lymphoid tissue. This model will be informed by detailed antibody data collected from individuals with a known infection history. As the project progresses I intend to consider the order of infection and vaccination, as well as which variants an individual is infected with. Further, the hypothesis of cross reactive antibodies already present in the population will be tested. The results of these individual level models will be used to inform population level models that aim to understand how repeat infection and vaccination protect the population from disease and infection. The intention of this analysis is to gain information about the longevity of protection from SARS-CoV-2 and to test hypothesis about whether, and in what form, it will be endemic. The novelty of this research stems from the methods being employed and their use of a wealth of available data. In the context of past research into SARS-CoV-2 infections no within-host model has considered the interplay between the competitive B-cell development in lymphoid tissue and the subsequent retrieval of virus by the antibodies the B-cells go onto produce. Further, the application of existing within-host models have mostly been to understand disease progression and risk of hospitalisation so focus damage to the host. As a consequence existing models often do not explicitly model immune process, but only their effects of virus and host cells, or if they do model immune process they are the ones that are related to cell damage, such as cytokine and cytotoxic t cells. Older models that are not specific to SARS-CoV-2 have tried to capture the process of the B-cell development. However, these models neglect the site of infection and make inaccurate assumption about the dynamics outside the lymphoid tissues. Further, these older models do not usually validate their conclusions with data. Where as, my parsimonious model captures both the competition in lymphoid tissue and relevant dynamics at the site of infection, and is observable and so can be validated by data. These properties make it more useful for understanding the effects of vaccination and informing the future vaccination campaigns.