Ivan Erill, Department of Biological Sciences
The understanding of promoter sequences is key to the further advance of genomics, since promoters constitute the fundamental blocks of transcriptional regulation, thus providing information about the transcriptional interactions among the genes they regulate. The analysis of promoter sequences is a complicated because promoters, as opposed to coding regions, show a great variability in their organization, structure and usage of the transcription factor alphabet of operators available in a cell. Traditional apporaches to promoter modeling are usually top-down and typically impose rigid rules on promorer structure that are not fitted for the general case, thereby restricting their application. In the light of this, soft-computing methods, such as artificial neural networks and genetic algorithms, seem clearly indicated to approach the modeling of promoters in a bottom-up, data-driven model.