My primary area of research is in Distributed Biomedical Systems, using Artificial Intelligence, in particular Semantic Web technologies, to address issues of integration and interoperability of computational and data resources across the Internet. An area of particular interest are Biomedical Atlases, where the combination of images and ontologies offer particular challenges and solutions. Here I am working in close collaboration with the MRC Human Genetics Unit’s Mouse Atlas project group.
Whilst in the past, most of this work has been based in the Life Sciences, more recently the focus is also shifting to include the clinical side. Many, though not all, the experiences gained in working with model organisms, such as mouse, apply to human as well. For example, the basic components of an atlas - the spatial (3D/4D) model, the symbolic representation (ontology), and the mappings between them - can be found in both domains. On the other hand, whilst accuracy, for example in the context of searching for genes expressed in a particular part of the heart in the mouse is important, but not absolutely critical, should some form of a brain atlas-based system be used for the planning of brain surgery, any recommendations to the surgeon better be highly accurate.
As Computer Science research generally is, or at least should be, driven by the eventual application of its outcomes in some domain, working in an interdisciplinary context is inevitable for those of us working closely with the relevant domain experts. However, even within Computer Science, a level of "interdisciplinary" research is required, as often it is the appropriate combination of CS techniques, rather than any single one, that holds the most promise.
For example, atlases combine work on image processing and ontologies. Over the next years, a major focus of my research interest will lie in what is generally referred to as the ”Big Data” challenge, particularly in the context of ever increasing amounts of spatio-temporal data in the biomedical domain, including clinical data such as from MRI or PET scans. Solutions will not only require to combine image processing and semantic web technologies, but bring to bear research in areas such as cloud computing, multicore processing and data mining.