Interpretation of two-dimensional electrophoresis gels
Résumé
The study of human protein polymorphism presents a major advantage in clinical chemistry to discover forms associated with risks or involved in well specified pathological states. If this polymorphism corresponds to structural differences reflected by physical property modifications it can be visualized on a map obtained by 2-dimensional electrophoretic technique. This study deals with apolipoproteins A-I, A-II, A-IV, C2, C3, and E that can be found on 2-dimensional gels. Each of these proteins appears on images under the form of a constellation of several spots. An expertise was extracted on these apolipoproteins in order to itemize all their known variants and their geometrical representations. This knowledge was implemented in a system to recognize what are the constellations we are actually dealing with, on a gel. A management system handles specific techniques taken from image processing to extract parameters from 2-dimensional gel images and artificial intelligence. Knowledge-based expert systems which perform the matching, allowing a partial interpretation of the gel. The accuracy of quantity measurement and quality control are greatly improved by adding calibration proteins. This same system will allow us to further implement conceptual clustering techniques to identify relevant relations in apolipoprotein metabolism. Apolipoprotein polymorphism is used in this work as a model which could be extended to other protein constellations in the future.