Use of average mutual information for studying changes in HIV populations

Abstract

Average mutual information (AMI) has been used in a number of applications in bioinformatics. In this paper we present its use to study genetic changes in populations; in particular populations of HIV viruses. Disease progression of HIV-1 infection in infants can be rapid resulting in death within the the first year, or slow, allowing the infant to survive beyond the first year. We study the development of rapid and slow progressing HIV population using AMI charts based on average mutual information among amino acids in the env gene from a population of 1142 clones derived from seven infants with slow progressing HIV-1 infection and four infants with rapidly progressing HIV-1 infection. The AMI charts indicate the relative homogeneity of the rapid progressor populations and the much greater heterogeneity of the slow progressor population, especially in later samples. The charts also show the distinct regions of covariation between residues without the need for aligning the sequences. By examining the changes in AMI between populations we can distinguish between clones obtained from rapid progressor and slow progressor. A measure of this change can be used to enhance prediction of disease progression.