Improving approximation of jitter probability in the breakpoints of simulated copy number alterations

Abstract

We analyze and approximate the jitter distribution function in the breakpoints of the measured genome copy number alterations (CNAs). The CNAs measured using the high resolution technologies of hybridization are contaminated with an intensive noise that may cause uncertainty in the detected breakpoints and segments. We show that jitter is fundamentally inherent to the simulate CNAs and that the jitter probability represented with the discrete skew Laplace distribution is not accurate when the signal-to-noise ratio (SNR) is small. To approximate the jitter distribution with highest accuracy, we modify the skew Laplace distribution to have the SNR function dependent on the discrete departure from the breakpoint. We propose several approximating functions and test them by experimental data.