Research Of Protein Spot Matching Algorithm For2-DE Images Based On Geometric Blocking And Gray Hierarchical

As one of the most important methods in the life science, the main work of two-dimensional gel electrophoresis (2-DE) is to extract differential protein spots from gels.Spot matching algorithm is a committed step of differential protein spots extraction,which may affect the accuracy of protein functionally identify and taxonomies directly.With the support of National Science Foundation and Jiangxi Science Foundation, basedon2-DE gel images, the study of auto spot matching method in different images is putforward. The main research content and achievements are as follows:1. The scientific background and analysis methods of proteomics were introduced.Common algorithm of spot matching in2-DE gel image was studied systematically anda brief analysis of research trends was given at the same time. Theoretical knowledgeand main process of image registration is elaborated.2. An auto-matching method based on local feature was presented. Firstly, Relativeneighborhood graph was built to measure local geometric feature of protein spots.Secondly, a feature space, which included central intensity and local geometry, wascombined with relative distance to measure similarities in coarse match. Thirdly,matched pairs in coarse matching were used to calculate coordinate transition model,the coordinate space of different image were kept consistent. Lastly, relations betweenresidual not matched points were measured by Euclid distance. The algorithm wasapplied to real2-DE gel images. The results showed that this algorithm had a goodmatching accuracy.3. To reduce the mis-matching or non-matching in2-DE images, an auto-matchingalgorithm based on gray hierarchical and geometric blocking was presented. Firstly,protein spots in the gel images were divided into groups by gray level and geometricposition, then a method based on shape context and normalized correlation was used tocoarse matching in protein spots. Secondly, matched pairs in coarse matching were setas feature points, the precise matching in the rest of not matched protein spots wereaccomplished by the method of geometric correlation and similarity criterion. Finally,local affine transformation was used in the verification of matching results to removenon-matching and mis-matching points. The algorithm was applied to different2-DEgel images. The results showed that the new matching algorithm could reduce the non-matching and mis-matching spots, and increase the matching accuracy.