Performance analysis and simulation of IIR anti-notch filter with various structures for gene prediction application

Abstract

Gene prediction is an important topic in genomic research. Various techniques are in use for identification of protein coding regions in genes. Application of Fourier technique is one of the most popular methods of gene prediction, in which prediction algorithm is based on period-3 property of DNA where it exhibits a prominent peak in coding region. Spectrum estimation by Fourier method generates various harmonics, generally known as 1/f noise along with sharp peaks, which may lead to false prediction of coding regions. Researchers used various parametric and non-parametric filters to tackle this problem and improve the accuracy of prediction. Performance of anti-notch filter with cascaded lattice structure on one hand and harmonic suppressor with comb filter on the other hand have been compared here for identification of coding regions of C-elegan F56F11.4a chromosome. The authors have analyzed the performance in terms of standard deviation and signal-to-noise ratio. A Matlab simulink environment has been used for filter realization and performance analysis.