Appropriate mathematical and statistical strategies can make sense of huge volumes of genomic data, aiding the discovery of gene function and the understanding of disease across a wide range of medical research.
Bioinformatics has been notoriously hard to define, despite the fact that over the years since the term has become common. Most attempts at definition fail by being too narrow or too broad. As further research over time goes on, however, it seems that the broadest definitions are the most helpful. One such is the following: “bioinformatics is the enabling discipline that stands at the intersection of the information and bio-technologies.” Although this definition does not specifically mention mathematics, statistics or computer science, or even data, it certainly brings to mind computers, and biological techniques. What else goes on inside computers but the storage and manipulation of data? And what else do we get from most biological techniques but data? After this is realised, the rest falls into place.
In our Division much of the data we deal with concern biological molecules – such as DNA, RNA, proteins and metabolites – as the narrower definitions of bioinformatics presuppose. However, we also have much to do with phenotypic data: observable characteristics of organisms, such as deafness status, survival time following cancer diagnosis, or drug sensitivity. We also deal with more abstract data such as familial relationships. Being in a medical research institute, one might think that all our phenotypic data would concern humans, but mouse data are everywhere, and we also get involved with marsupial (opossum, wallaby), monotreme (platypus), and other primate (chimp, gorilla, monkey) data. Not only are these non-human organisms interesting in themselves, their study tells us a lot about humans, as we can exploit evolution, particularly at the molecular level.
The geneticist Dobzhansky famously contended in 1964 (and later in 1973) that “Nothing in Biology Makes Sense Except in the Light of Evolution”. Molecular evolution is one of the dominant themes in bioinformatics, and Dobzhansky’s aphorism could well be rephrased “everything in biology benefits from exposure to the light of evolution.” We use mouse models and comparative genomics, which clearly build on evolutionary relationships, but less obvious, and in a way more fundamental, every database search we do, and every sequence alignment we make uses algorithms based on evolution. Keep this in mind as you read the sampling of our research that follows.
Professor Gordon Smyth (Division Head)
For further information go to the Bioinformatics website which has software, seminar listings and more