Manifestation and exploitation of invariants in bioinformatics

Abstract

The accumulation of huge amount of biomedical data and the need to turn such data into useful knowledge lead to many challenging bioinformatics problems. Many techniques have been developed for the bioinformatics problems that have emerged, and more are being proposed everyday. I present here a selection of these problems and techniques, highlighting a fundamental property that is common to all of them. Specifically, I observe that these problems are characterized by invariants that emerge naturally from the causes and/or effects of these problems, and show that the techniques for their solutions are essentially exploitation of these invariants. In the process, we learn several major paradigms (invariants, emerging patterns, guilt by association), some important applications (active sites, key mutations, origin of species, protein functions, disease diagnosis), some interesting technologies (sequence comparison, multiple alignment, machine learning, signal processing, microarrays), and the economics of bioinformatics. © 2008 Springer-Verlag Berlin Heidelberg.