Automatic registration for images of two-dimensional protein gels

Two-dimensional polyacrylamide gel electrophoresis 1 (2-D PAGE 1) is currently the method of choice for separating complex mixtures of cellular proteins. Despite its usefulness, 2-D PAGE is not being applied to its full potential because of difficulties with both the method and analysis of the results. One of the key problems is the difficulty and slowness of image analysis, especially registration (image alignment), which is laborious and the results unsatisfactory. We have developed a novel system for analysis of 2-D PAGE images, called Z3, that performs the analysis faster and more precisely. The Z3 system employs novel approaches to image registration, image display, computation of differential expression, and the design and analysis of 2-D gel experiments. This paper describes in detail the registration algorithm, and briefly discusses the merits of complementary pseudocolor display. The registration algorithm is novel in that for the first time raw-image-based registration technology is applied to 2-D gel analysis.