Two-part permutation tests for DNA methylation and microarray data

The proposed permutation test leads to smaller p-values in comparison to the original two-part test. We found this for both DNA methylation data and microarray data. With a simulation study we confirmed this result and could show that the two-part permutation test is, on average, more powerful. The new test also reduces, without any loss of power, to a standard test when there are no null or truncated values.The two-part permutation test can be used in routine analyses since it reduces to a standard test when there are positive values only. Further advantages of the new test are that it opens the possibility to use other test statistics to construct the two-part test and that it avoids the use of any asymptotic distribution. The latter advantage is particularly important for the analysis of microarrays since sample sizes are usually small.The addition of a methyl group at the carbon-5 position of cytosine is a modification of DNA called DNA methylation. In mammalian cells, DNA methylation is essential for proper development [1]. The methylation patterns of tumor cells are altered compared to those of normal cells, moreover, there are also differences between different types of cancer as shown for subtypes of leukemia [2] and lung cancer [3]. Thus, DNA methylation analysis promises to become a powerful tool in cancer diagnosis [4].DNA methylation data can be obtained using the MethyLight technology [5]. When the tested region is not or only partially methylated the result is negative (undetectable methylation, null values). In contrast, samples that show methylation will have a value greater than 0 [4]. Thus, DNA methylation data obtained with MethyLight have a clump of zero observations and a continuous nonzero part. For such a data structure, two-part models as proposed by Lachenbruch [6-8] are applicable. In that approach, the test statistic is the sum of two squared statistics, one comparing the proportions of zeros and one comparing the positive values. For exam