Abstract
We present a method for clustering genomic sequences based on variations in local entropy. We have analyzed the distributions of the block entropies of viruses and plant genomes. A distinct pattern for viruses and plant genomes is observed. These distributions, which describe the local entropic variability of the genomes, are used for clustering the genomes based on the Jensen-Shannon (JS) distances. The analysis of the JS distances between all genomes that infect the chlorella algae shows the host specificity of the viruses. We illustrate the efficacy of this entropy-based clustering technique by the segregation of plant and virus genomes into separate bins.
Highlights
The organization of genomes has evolved dynamically by a stochastic process comprised of mutation and selection
We show that the local variations in entropy are very useful for clustering viruses and plant genomes, but may suggest host specificity of viruses
We have proposed a novel method for clustering genomic sequences based on variations in local entropy
Summary
The organization of genomes has evolved dynamically by a stochastic process comprised of mutation and selection. When applied at the nucleotide level to genomic sequences, the sequence entropy for the full genome can be reduced to the question of CG-content [1,2]. This is a global property and related to the mutation rate, while the selective advantage reveals itself locally in, e.g., gene products such as regions coding for proteins.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have