Identification of CpG islands in the human genome through Markov chains: A probability-based mathematical model.

Abstract

The biologically important genes known as essential genes or housekeeping genes are usually found surrounded by regions called "CpG Isles" . The CpG isles are named so because they contain a much larger quantity of dionucleotides CpG than the rest of the genome. Since during the recognition of such isles the location of the housekeeping genes can be inferred, a mathematical model of identification of CpG isles will make it easier to tell it apart from the rest of the genome. The mathematical model that is presented in this article uses as an example a sequence of 60 nuc1eotides present in the genome of the canine parovirus and is based on the Markov chains to calculate the probability that a fragment of this sequence. in relationship to the rest of it, corresponds or not to a CpG isle. This model can be used in any sequence, independently from its number of nucleotides. However the parovirus sequence, chosen in this case as a small sample. served to compare and confirm the results by simple inspection