Identification of CpG islands using a bank of IIR lowpass filters

Abstract

It has been known that biological sequences such as the DNA sequence display different kinds of patterns depending on their biological functions. This statistical difference can be exploited for identifying the region of interest, such as the protein coding regions or CpG islands, in a new biological sequence that has not been annotated yet. A region of particular interest is the CpG island, which is a region in a DNA sequence that is rich in the dinucleotide CpG, since it is known that they can be used as gene markers. There have been several computational methods for identifying CpG islands, each with its own strength and weakness. In this paper, we propose a novel scheme for detecting CpG islands in a genomic sequence, which is based on a bank of IIR lowpass filters. The proposed method is capable of identifying CpG islands efficiently at a low computational expense. Simulation results are included where appropriate to demonstrate the idea.