Whilea general solution for causal inference based solely on conditional independencies has been knownfor a while (the celebrated "docalculus"of Judea Pearl), in practice conditional independencies alone are far fromenough. For instance, if hidden common causes confound the relationship between causes and effects, then only in very specific cases graphical information can help. In contrast, the idea of using mathematical programming to find bounds on causal effects has not achieved much traction due to its computational cost and limited waysof expressing soft assumptions that complement graphical structures. This PhD thesis proposes an expansion ofthemathematical programming formulation that will exploit advances in stochastic optimisation and message passing. This "causal programming"formulation aims to provide anew foundation for (partially) identifying causal effects, hopefullykickstarting a new generation of tools for causal inference.

When vital organs are damaged in disease this results in the development of scar tissue. When the scar is excessive or disporportionate this can itself worsen organ function, a process termed fibrosis. Fibrosis is a result of altered or abnormal healing processes and is a common factor in many common diseases. Less common disorders such as scleroderma (also called systemic sclerosis) cause fibrois in multiple organs including the skin, lungs, heart and blood vessels. We have developed, for the first time, models of fibrotic disease that can be used to study the development and consequences of fibrosis in the skin and internal organs. Through understanding these models, it is very likely that better methods for assessment and treating fibrosis will emerge.

Hereditary retinal disease affects our most precious sense, sight, in about 1 in 3000 of the population. Many of these disorders cause complete loss of vision blindness and, unfortunately, there is as yet no effective treatment. A major difficulty in developing treatments is the inability to sample retinal tissues from patients in order to fully understand how the disease develops or to test novel treatments that may further compromise limited vision. The use of animal models avoids these problems but, to be meaningful, they need to accurately mimic the human disorder. This means that the same gene mutation must be studied. Such mutations are generally not found in nature so the same gene as in human must be targeted in the animal so that the same mutation is introduced. These are complex, time-consuming and expensive experiments, so it is important that such models of human disease are fully exploited. We have already generated two such models with human mutations in genes responsible for basic visual processes resulting in the loss of light sensitive cells in the retina. We propose to use these models to gain a detailed understanding of the process of retinal degeneration and to follow this with an assessment of potential treatment options that will include the testing of various drugs, growth factors and gene replacement therapy using viral vectors delivered to the eye. This will enable us to determine the treatment for patients and will provide the first step towards a clinical trial.

Aging is accompanied by a marked susceptibility to infectious diseases, which inflict heavy toll upon the rapidly aging society with regard to lost productivity, mounting health costs and loss of life. The progressive decline with age of the immune system - immunosenescence - is the primary underlying cause of the age-related increase in this susceptibility. Despite decades of research, important gaps remain in our understanding of the fundamental nature of the process, as well as in our practical ability to protect older adults against infectious diseases. Our bodies mount immune responses when attacked by infectious organisms. The central participants in this response are our white cells (leucocytes), whose role is to protect us from the infective agents. Based on the molecules they secrete different populations (subsets) of leucocytes have been identified which play different roles during immune responses. A separate subset of leucocytes is known as regulatory cells and their main role is the control of other leucocytes. The major problems that are inherent in the study of human immunology is that generally, only leucocytes from the blood compartment are studied and that virtually all studies are performed in vitro. We have developed and validated a new model for the study of a human memory T cell response in humans in vivo. This involves the injection of a recall antigen in the skin, followed by either the biopsy of the lesion for histological analysis of the immune infiltration, or the harvest of infiltrating cells using suction blister technology. Using this model, we propose to clarify the functional profile of cells that accumulate in the skin during immune responses induced by different antigens and determine whether these leucocyte populations behave differently in the old individuals. Specifically we wish to examine the relationship between regulatory cells and other functional leucocyte populations and explore mechanisms which control their behaviour. This information may enable us to manipulate this balance artificially.

Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.