ABSTRACT:

Alterations in the number of DNA copies of portions of a genome are both a process that contributes to natural genetic diversity in healthy individuals and also a driving cause for cancer and other genetic diseases. Even though recent advances in microarray technology and high throughput sequencing now allow very high resolution scans for very large cohorts of samples, small alterations remain particularly difficult to detect, especially under severe noise degradation conditions. Thus, estimation of the DNA copy number remains a challenging problem. Some of the core challenges in copy number estimation can be seen as essentially signal processing problems, where it is necessary to exploit fundamentally different characteristics between the desired underlying biological signal and the measurement noise. Indeed, some of the most successful methods used to address this problem have been inspired by well-known signal processing techniques.