Educating biologists in the 21st century: bioinformatics scientists versus bioinformatics technicians

For many years algorithms were taught exclusively to computer scientists, with relatively few students from other disciplines attending algorithms courses. A biology student in an algorithms class would be a surprising and unlikely (though not entirely unwelcome) guest in the 1990s. Things change; some biology students now take some sort of Algorithms 101. At the same time, curious computer science students often take Genetics 101. Here comes an important question of how to teach bioinformatics in the 21st century. Will we teach bioinformatics to future biology students as a collection of cookbook style recipes or as a computational science that first explain ideas and builds on applications afterwards? This is particularly important at the time when bioinformatics courses may soon become required for all graduate biology students in leading universities. Not to mention that some universities have already started undergraduate bioinformatics programs and the discussions are underway about adding new computational courses to the standard undergraduate biology curriculum a dramatic paradigm shift in biology education. Since bioinformatics is a computational science, a bioinformatics course should strive to present the principles and ideas that drive an algorithm’s design or explain a crux of a statistical approach, rather than to be a stamp-collection of the algorithms and statistical techniques themselves. Many existing bioinformatics books and courses reduce bioinformatics to a compendium of computational protocols without even trying to explain the computational ideas that drove the development of bioinformatics in the last 30 years. Other books (written by computer scientists for computer scientists) try to explain bioinformatics ideas at the level that is well above the computational level of most biologists. These books often fail to connect the computational ideas and applications thus reducing a biologist’s motivation to invest time and effort into such a book. We feel that focusing on ideas has more intellectual value and represents a long-term investment: protocols change quickly, but the computational ideas don’t seem to. However, the question of how to deliver these ideas