Copy number detection using self-weighted least square regression

In this article, an efficient algorithm to detect the breakpoints in DNA copy number alterations is considered. In view of the influence of the heavy noises, the self-weighted least square estimation is adopted to downweight the covariance matrix of the wild observations (outliers), which ensure the convergence between the estimated parameters and the true values. The proposed approach makes use of the most of the data itself to reduces the complexity of the model, and presents an insightful discussion for discovery of copy number alterations.