How Does Computer Science Change Molecular Biology?

Abstract

One of the most frequent opinion about the advent of ‘computational biology’ states that the flood of data emerging from the genomic and post-genomic projects created a growing demand of computational resources for their management and analysis thus giving rise to an interdisciplinary field between computer sciences and molecular biology. This point of view is not completely satisfactory. First, it precludes the historical background of a field that appears, in the 70’s, at a time where the sequence data were not flooding at all. Second, it emphasizes the quantitative aspect of the data produced by (new) experimental technologies whereas the real difficulties probably rest in the diversity and complexity of the biological data. For instance, all the human chromosomes totalize about 3 Gb in size, which is not such a large quantity of information (by comparison to, let’s say, meteorological records). By contrast, the internal structure of these chromosomes such as the location of gene or regulatory sites or their large-scale organization remains a huge puzzle. This situation is still more complex in the so-called ‘post-genomic’ area. The biological concepts involved in the analysis of regulatory or metabolic networks, for instance, ask for much more sophisticated representations than sequence of symbols. Third, and most importantly, this statement gives to computer science and biology two distinct roles : the latter providing (nice) problems to the former that, in turn, is supposed to solve them. Unfortunately, the situation is not so clear-cut. First, it hardly ever appends that biology can provide a problem under a form directly suitable for computer analysis.