Genetics of cancer risk and therapy response
The lifetime risk of cancer is estimated to be about 50%, but this risk is not the same for every person. It is well-established that the risk of cancer is influenced by inheritable genetic variants as well as lifestyle factors such as smoking and also medical conditions including obesity and diabetes. It is not clear, however, how these different strands of information, which may at times contain both favourable and unfavourable indications, can best be combined to measure each person's individual cancer risk. In this study proposal we use and develop computer algorithms to detect, quantify and combine cancer risk factors from the wide array of genetic, medical and lifestyle information provided by participants of the UK Biobank. These analyses may reveal novel cancer risk associations, but most importantly enable calculating the combined risk of developing cancer based on genetics, medical history and lifestyle factors. A secondary aim of this study, following the same approach, is to assess how genetics and medical history influence cancer outcomes. The intended duration of this project, which is based on the analysis of existing UK Biobank data, is 36 months. The ability to measure inter-individual differences in cancer risk is important for implementing effective cancer screening programmes, because high-risk individuals should be tested more thoroughly, while low-risk individuals may be harmed by unnecessary tests and treatment. Similarly, this information may facilitate a timely cancer diagnosis, because early symptoms are often unspecific, and the interpretation of such symptoms may change knowing that an individual has a higher or lower cancer risk. Understanding the factors influencing cancer therapy outcomes may help choose the most effective treatment for a cancer patient.
Scope extension: The research question addressed in this proposal is how accurately cancer risk can be calculated from inherited genetic variants, lifestyle, and health data. Epidemiological research has shown that about 10-20% of cancer risk can be attributed to heritable factors and that another 40% of cancers can be explained by lifestyle factors. However, it is unclear how these different and highly complex layers of information are best combined into a single figure representing an individual's risk to develop cancer in a given time frame. Here we will use statistical machine learning algorithms to develop integrative cancer risk prediction models, based on genetic variants, health and lifestyle information, and assess how accurately these models quantify each individual's cancer risk and which elements of information are most predictive.
Additionally, we want to explore an individual's risk for related comorbidities, precursors, and competing diseases. A particular focus will be on Diabetes, Dementia, and Cardiovascular disease.