This calibration analysis gives us the ability to perform a formal power analysis of our OB-HMM method (Figure 2). We believe this to be a fundamental reversal in the normal practice of developing CNV detection algorithms, where algorithms are first developed, then applied to the experimental sample, and false positive and false negative rates subsequently inferred via independent experimental validation of detected and non-detected CNV events. In our strategy, the characteristics of the algorithms are defined before application to the experimental sample, via the calibration analysis, offering the experimentalist the capability of being able to specify the desired false positive rate suitable for their experiment a priori.
