Background: The genomic sequences can be expressed in terms of alphabets and hence
they are discrete in nature. Therefore, digital techniques to analyze genetic problems are in need.
Objective: The main aim is to use digital signal processing techniques for the detection of CpG
Method: A method to detect the CpG islands using wavelet filtering has been proposed. Modified
Electron Ion Interaction Potential mapping has been proposed for numeric conversion. The signal is
restricted in frequency through a band pass filter and then wavelet filtered. CpG islands produce
bigger magnitude coefficients in the wavelet domain.
Results: The proposed method has been tested on genomes of Homosapiens, Saccharomyces
cerevisiae, Drosophila melanogaster, Caenorhabditis elegans, Escherichia coli, Zebrafish, Arabidopsis
thaliana and Mus musculus, downloaded from the national center for biotechnology information
database. Standard performance metrics have been evaluated and the values obtained are sensitivity -
84.68%, specificity – 85.6%, accuracy – 82.22% and correlation coefficient – 63.31%.
Conclusion: On comparing the evaluation metrics obtained to the methods in the literature, it is found
that the wavelet transformation method is better. The area under the receiver operating characteristic
curve has also been evaluated and is 0.8705 which is larger compared to the methods in literature.
Hence, it can be concluded that the proposed method is efficient in detecting CpG islands.