Detecting genetic variants of breast cancer using different power spectrum methods

Abstract

Cancer is one of the most dangerous diseases that the world faces and has raised the death rate in recent years, it is known medically as malignant neoplasm, so detection of it at the early stage can yield a promising approach to determine and take actions to treat with this risk. Bioinformatics is the application of computer technology to the management of biological information. Its field has now solidly settled itself as a control in molecular biology and incorporates an extensive variety of branches of knowledge from structural biology, genomics to gene expression studies. Genomic signal processing (GSP) techniques have been connected most all around in bioinformatics and will keep on assuming an essential part in the investigation of biomedical issues. It refers to using the digital signal processing (DSP) methods for genomic data (e.g. DNA sequences) analysis. Recently, GSP is one of important methods which can be used to detect the cancerous cells that are often caused due to genetic abnormality. In this paper, some of GSP techniques that depend on frequency domain transformation are presented. These techniques are: Discrete Fourier Transform (DFT), Power Spectral Density (PSD) of DFT and PSD obtained by Welch's averaged periodogram method. The proposed method has given satisfied results for differentiation between normal and cancerous cells. The algorithm is tested on six healthy and six cancerous genes of breast cell which are obtained from NCBI genbank.