Examining coding structure and redundancy in DNA. How does DNA protect itself from life's uncertainty?

0739-5175/06/$20.00©2006IEEE T he genetic code instructs proteins on the translation of nucleotides to amino acids, but this example is only one of many signals encoded in DNA. It is well known that these protein-coding regions have the lowest mutation rates in the DNA strand. So, the question arises: how does DNA protect itself from error? A review of DNA signal content, redundancy, and mutational mechanisms is presented. Then, mutation-robust methods are developed to detect a linear coding structure and approximate tandem repeats. Ever since the introduction of the Watson-Crick model of DNA, scientists have been trying to decipher the long sequence of millions (or, for complex organisms, billions) of bases. The genetic code, the mapping of nucleotide triplets (codons) to amino acids, or “protein-coding,” was one of the first discoveries. Signals in DNA could then be paralleled to digital signals. After 30 years, many functions and signals in DNA still remain unknown, and scientists have conjectured that nonprotein-coding regions, which compose 97% of human DNA, are unused junk [1]. On the contrary, recent studies reveal that binding sites and initiation signals exist in these nonprotein-coding areas, and mutation errors in these regions cause diseases [2]. Nonprotein-coding regions contain a finite amount of “algorithmic” content [1], [3]. Discovering the signals and function in these areas is just the beginning of genome discovery. In this article, we develop a method to uncover an errorcorrection coding structure in the nucleotide sequence, and show that our framework is efficient for detecting approximate tandem repeats, such as microsatellite regions.