Recent advances in biotechnology, spurred by the Human
Genome Project, have resulted in the accumulation of vast amounts of new
data. Ontologies -- computer-readable, precise formulations of concepts (and
the relationship among them) in a given field -- are a critical framework for
coping with the exponential growth of valuable biological data generated by
high-output technologies. This book introduces the key concepts and
applications of ontologies and ontology languages in bioinformatics and will be
an essential guide for bioinformaticists, computer scientists, and life science
The three parts of Ontologies for Bioinformatics ask, and answer, three pivotal questions: what ontologies are; how ontologies are used; and what ontologies could be (which focuses on how ontologies could be used for reasoning with uncertainty). The authors first introduce the notion of an ontology, from hierarchically organized ontologies to more general network organizations, and survey the best-known ontologies in biology and medicine. They show how to construct and use ontologies, classifying uses into three categories: querying, viewing, and transforming data to serve diverse purposes. Contrasting deductive, or Boolean, logic with inductive reasoning, they describe the goal of a synthesis that supports both styles of reasoning. They discuss Bayesian networks as a way of expressing uncertainty, describe data fusion, and propose that the World Wide Web can be extended to support reasoning with uncertainty. They call this inductive reasoning web the Bayesian web.
Kenneth Baclawski is Associate Professor of Computer Science at Northeastern University.
Tianhua Niu is Assistant Professor of Medicine at Harvard Medical School and Director of Bioinformatics, Division of Preventive Medicine, at Brigham and Women's Hospital, Boston.
Slide presentation at the CSB2005 Conference
Updated list of URLs in the textbook. I check the validity of these links regularly, but it is hard to catch all of the changes. If you find that one of these links is incorrect or no longer valid, please send me email at Ken@Baclawski.com.
The textbook omitted the multiplicity for nitrogen in nitrous oxide (N2O). This mistake illustrates a deep and important incompatibility between traditional data modeling as in XML and relational databases, and the logic used by the Semantic Web. Nitrous oxide and the problem of default attributes introduces some interesting topics that are motivated by this incompatibility and that go beyond the scope of the textbook.