Modelling gene regulatory data using artificial neural networks

The amount of biological data has increased dramatically in recent years and computational models have struggled to keep up with this trend. This is especially the case in microarray data which plots the activity of genes within a cell. The gene activities within a cell are regulated by other genes within that cell, and it is the extraction of these patterns (or regulatory networks) within the data which is of great interest to biologists and computer scientists alike. We describe experiments using artificial neural networks to assimilate the microarray data and construct gene regulatory networks. The resulting networks are capable of encoding complex relationships between genes and to correctly predict many cell states given only the starting state of the cell.