Methyl-IT implements methylation analysis based on signal detection theory and machine-learning. Methylation changes are expressed in terms of information divergence (ID) of methylation level. Methylation process is a stochastic process, particularly, a biochemical-biophysical process which must not be reduced to statistic. Methyl-IT includes the information on the statistical biophysics of the methylation process with the estimation of the probability distribution of the methylation background noise (plus signal), which is used as null distribution for the application of basic signal detection theory.