Maximum likelihood principle for DNA copy number analysis

Microarray technologies had been used to measure DNA copy number data. The copy number represents the relative fluorescent intensity level between control and test DNA samples. Variation of this number may lead to many genetic diseases such as cancer. Unfortunately, the observed copy numbers are corrupted by noise due to experimental errors and probes accuracy, making the variations hard to detect. Different techniques had been proposed to denoise the data and to extract the important feature such as the breakpoints from the variant regions. In this paper, we present a robust procedure for the analysis of DNA copy number data based on maximum likelihood principle using global information of the entire data record. We show that Dynamic programming can be used to compute the DNA copy number estimates and reduce the computational complexity. Furthermore, we employ the Minimum Description Length rule to estimate the number of unknown parameters. Using simulated and real data, we show that the proposed method outperforms other popular commercial software and published algorithms.