Bioinformatics has been evolving for the last 20 years. One of the major thrusts of the field is inferring the functions of proteins. There are many techniques for this: sequence similarities, common structural motifs, gene order, gene fusion events and similarities in gene expression. This gives a static picture of the protein function in a biochemical pathway. However, biological systems are, by their nature, dynamic. As a consequence, the focus is changing. Mathematical and computational methods are being applied to predict function based on the role of genes and proteins in networks. These methods are old; the increase of computing power and high-throughput biological data makes their application more viable now.