Probabilistic bounds for estimates of genome DNA copy number variations using HR-CGH microarray

Estimation of the genome copy number variations (CNVs) measured using the high-resolution array-comparative genomic hybridization (HR-CGH) microarray is provided in the presence of large measurement white Gaussian noise having typically different segmental variances. Jitter inherent to the breakpoints of such signals can be approximated with the discrete skew Laplace distribution. Referring to and aimed at sketching a more clear picture about possible chromosomal changes, we have justified the estimate UB and LB probabilistic masks in the three-sigma sense to guarantee an existence of true changes with the probability of 99.73%. Some real measurements are tested by these mask and practical conclusions are provided.