PATTERN ANALYSIS OF BIOLOGICAL SEQUENCES.

Abstract

Some typical types of analyses of biological sequence derived from macromolecules such as nucleic acids (i. e. , deoxyribonucleic acid or DNA and ribonucleic acid or RNA) include the detection of global dependencies and the assessment of the coding potential over different regions of the sequence. The relative entropy measure is shown to be directly applicable to these two types of problems encountered in DNA sequence analysis. Furthermore, a statistic based on the relative entropy value exists and can be used for determining the significance of observed values. The appropriate use of the relative entropy measure for each type of analysis is discussed in detail and applied to several real DNA sequences from the bacteria Escherichia coli (E. coli), each of which is many thousands of bases in length. The methods developed are not restricted to biological sequences but pertain in general to the analysis of sequences composed of characters.