Model Based Protein Lysate Arrays Data Analysis

Despite the enormic genomic complexity of most organisms,and in particular humans,the complexity is further increased at the protein level as a result of posttranslational modifications ,such as phosphrylation ,acetylation and ubiquitination,which can appreciably impact the function state of proteins. Therefore, it is necessary that protein concentration and the changes of protein concentration are studied, and thus can make more in-depth understanding of biological systems. The data of protein lysate microarray of all samples and successful dilution of observed-expression are known, based on these data, now there are four ways to build different models, obtained relationship between observed-expression and the true protein concentration. The first three methods are Linear-Modeling(Micrean,C.et al.,2005),Nonlinear-Modeling(Tabus,I.et al.,2006),Non-parametric(Hu,J.et al.,2007), They are to establish the curve between Observed-expression and protein concentration, just the curve is in different type. The last method is Serial dilution curve(Zhang,L.et al.,2009).It is still trying to find the relationship between observed-expression and protein concentration,and has specific ideas ,and is different from the previous three methods, this article will explain the details. These four methods have advantages and disadvantages, the ultimate goal is to calculate the initial concentration of protein based on data.In this article, these four methods make simulation in different model and compare the simulation results, the final to determine the scope of application of various methods.