Detecting Regulator-Target Gene Pairs from Expression Profile of Microarray

Generally time-series microarrays are widely used to detect the regulation structure because they contain useful information about the relationship among a few thousand genes. Lots of different approaches have been introduced to detect regulator-target pairs using whole expression data. In this paper, we present a simple, but efficient method for detecting the regulator-target pairs based on an alignment method. Comparing to previous work based on event-string comparison approach, the basic idea of our algorithm is different in that we have applied a parametric optimization for alignment scoring matrix by using a simple machine learning procedure. As a result, we get a sensitivity and a specificity that is 10% higher than the recent work by Kwon (2003).