Image analysis

Owing to the volume of data and its experimental geometric and expression uncertainties, quantitative analysis of 2D gel electrophoresis data with image processing and modeling is an actively pursued research topic. The results of this analysis include accurate protein quantification, pI and M[[inf]]r[[/inf]] estimation, and the extraction of differential expression between control and sample gels. Systematic errors such as current leakage and background inhomogeneities need to be corrected for, followed by the segmentation, modeling, and quantification of each protein spot. To establish spot correspondences between multiple gels, a number of image registration and spot matching techniques for correcting geometric distortion have been proposed. They have facilitated the detection and quantification of differential expression through the statistical analysis of outliers and bias field. For large-scale analysis, classification and clustering techniques can be employed to uncover intrinsic trends and relationships in the protein spot patterns. This section provides a short review of the computational techniques used in the analysis of 2-DE gels and examines the pitfalls of existing methods. Some of the key areas that need to be developed in the coming years are highlighted, especially those related to statistical expression analysis (SEA) based on multiple gel runs, and image mining techniques through the use of high-throughput processing based on the Grid technology.