Current genomic microarray technology has become an advanced testing procedure for many different fields related with or supported by Functional Genomics, Microarray technology usage has experienced a certain explosion in the past years, well defined as a gold rush in a parallel metaphor. This technology, conceived initially as a parallel implementation of well known techniques as Northern and Southern Blotting, has become an available and easy-to-use standard procedure for the estimation of gene expression levels. Nevertheless important challenges have still to be faced to provide it with the desirable reliability levels required for its proper use. Many are the factors which result in unreliable estimates of expression levels, which are to be solved more by the microarray engineer or statistician than by the expert in Genomics. These are known as microarray data processing challenges. The importance of side fields of knowledge as Signal and Image Processing, Pattern Recognition, Statistical Data Analysis based on Independent Component Analysis in relation with microarray data processing challenges have not completely yielded their enormous potential in solving problems as microarray image enhancement, segmentation, correction, gridding, data analysis, reliable expression estimation in relation with hybridization dynamics, etc.
The thesis will concentrate on determining genes showing differences in their expression levels, differences considered to be the effect of underlying processes affecting the hybridization on its dynamic evolution. Independent Component Analysis applied to these genes may be the tool which could give a solution to this issue together with hybridization modelling.