2.Methods for Identification DMRs
2.1  DMRs based on C

Static window method

In this method, the window length and the number of adjacent windows are fixed. Users could set the above two parameters on the web page. The wilcoxon test will be conducted if both samples have sufficient coverage in these windows and the C methylation level of one sample is at least 0.2 (optional parameter) greater than the other. The p-value and the FDR correction is used and determined by users, and the test window moves one methylcytosine at each time. Then the DMRs with short distance are joined together.

Dynamic window method

In this method, the base C number is fixed in seed (optional parameter). The wilcoxon test will be conducted if the percent of C coverage is sufficient (optional parameter) for both samples and the methylation level of one sample is at least 0.2 (optional parameter) greater than the other. The window will move a step size (C number, optional parameter) to the 3�� direction once a time, and repeat the wilcoxon test until the p value is not significant or to the end of the genome; Then, this window will move a step size (C number, optional parameter) to the 5�� direction and repeat the wilcoxon test until the p value is not significant or to the beginning of the genome. Finally all the districts searched by this method will be joined together if there is overlap.

2.2  DMRs based on CG

Static window method

In this method, the window length and the number of adjacent windows are fixed. Users could set the above two parameters on the web page. The wilcoxon test will be conducted if both samples have sufficient coverage in these windows and the CG methylation level of one sample is at least 0.2 (optional parameter) greater than the other. The p-value and the FDR correction is used and determined by users, and the test window moves one methylcytosine at each time. Then the DMRs with short distance are joined together.

Dynamic window method

In this method, the base CG number is fixed in seed (optional parameter). The wilcoxon test will be conducted if the percent of CG coverage is sufficient (optional parameter) for both samples and the methylation level of one sample is at least 0.2 (optional parameter) greater than the other. The window will move a step size (CG number, optional parameter) to the 3�� direction once a time, and repeat the wilcoxon test until the p value is not significant or to the end of the genome; Then, this window will move a step size (CG number, optional parameter) to the 5�� direction and repeat the wilcoxon test until the p value is not significant or to the beginning of the genome. Finally all the districts searched by this method will be joined together if there is overlap.

2.3  DMRs based on CH

Static window method

In this method, the window length (optional parameter) and the number of adjacent windows (optional parameter) are fixed. Users could set the above two parameters on the web page. The wilcoxon test will be conducted if both samples have sufficient coverage (optional parameter) in these windows and the CH methylation level of one sample is at least 0.2 (optional parameter) greater than the other. The p-value and the FDR correction is used and determined by users, and the test window moves one methylcytosine at each time. Then the DMRs with short distance are joined together.