Towards the genome clinic of the future - new computational methods to interpret genome variation and diseases

All human beings are approximately 99.9% similar in their DNA sequences. The 0.1% contains different DNA variants that contribute to the various phenotypic differences, such as different height and BMI. While 0.1% of the genome does not seem like an impressive number, this means millions of variants can be found in a typical healthy individual. To date, people have observed tens of millions of DNA variations and most of them do not cause diseases. However, it is challenging to distinguish between the pathogenic variants and benign variants, and it is often difficult to interpret the molecular consequences of genetic variants. This thesis presented several computational methods and tools to help with variant interpretation. Chapter 1 introduced the state-of-art variant interpretation methods and discussed current challenges for variant interpretation. Chapter 2 presented a pathogenicity estimation tool for rare variant interpretation. Chapter 3 explored potential protective genetic and environment factors on a rare pathogenic variant for dilated cardiomyopathy. Chapter 4 showed that co-methylation patterns are complementary to co-expression patterns for gene function interpretation. Chapter 5 utilised scRNA-seq data and identified variants influencing cell-type-specific gene-regulatory patterns. Chapter 6 reviewed the proposed computational methods and discussed potential future research directions on variant interpretation and their applications in the clinic