The Application Of The R Language Package InfiniumPurify In Tumor Purity Estimation And Differential Methylation Analysis

DNA methylation is closely related to tumor diseases and human development.Estimation of tumor cell purity and differential methylation analysis are important aspects in epigenetics research.However,based on DNA methylation chip data to study the purity of tumor cells and differential methylation analysis methods are not comprehensive.In particular,because of the tumor samples of we obtained are mixed tumor-normal samples,the tumor purity will cause a deviation from the differential methylation analysis and even to produce erroneous results.The existing method has not yet fully achieved the "correction" of tumor samples,and there is no method of estimating the methylation level of pure tumor samples currently.In view of the above problems,this paper presents the InfiniumPurify package,which including the following three models: First,the getPurity function to estimate purities,which based on the beta value matrix to get the most differentially methylated probes(iDMCs),then methylation levels of these selected iDMCs are used to vote for tumor purity after hyper-/hypo-transformation.The results of the prediction are highly correlated with those from ABSOLUTE and other available tools.Second,InfiniumDMC performs differential methylation analysis according to the tumor purity.Because of according to the purity avoiding the presence of errors which caused by impure of tumor samples.The method provides more sensitive results compared with existing methods.InfiniumDMC also performs differential methylation analysis without normal samples,greatly expanding the TCGA data analysis and application;Third,InfiniumPurify function is to deconvolute pure tumor cell methylomes from tumor samples,normal samples and tumor purity through a linear regression model.After we corrected the purity its difference between purified tumor and normal methylomes is very significant.