Study On Differential Methylation Analysis Methods Of Tumor Samples With Uneven Tumor Purity

This paper mainly studies on the detection of differential methylation sites in two groups of tumor samples.DNA methylation is an important epigenetic modification that plays an important role in cellular processes(including gene regulation,development,and disease),and in dysregulations are widespread in most types of cancer.Inferred differences methylated Cp G sites between two groups of tumor samples with different genotypes or phenotypes is a key step to reveal the epigenetic mechanisms of tumorigenesis,and identify biomarkers cancer subtypes.However,tumor purity as the main source of confounding factors,uneven tumor purity distribution between the two groups of tumors samples will lead to deviations in differential methylation sites if not properly considered.We first proved that the effect of tumor purity on differential methylation analysis is multiplicative,and then we propose Infinium DM,a generalized least squares model for adjusting tumor purity effect in differential methylation analysis.Our method is applicable to a variety of experimental designs,including with or without normal controls,different sources of normal tissue contaminations.By using simulated dataset,we compared our method with traditional methods including minfi,limma and limma corrected with tumor purity.Our method clearly shows significantly better performance under different scenarios such as different levels of differential methylation threshold,sample sizes,mean purity deviations,etc.,We also performed a series of verifications on real data to evaluate the performance of our proposed method.Overall,simulation and real data analysis demonstrate that our proposed method shows favorable performance over existing methods with similar purpose.