Coding sequence density estimation via topological pressure (Q2512920)

In this paper, the authors present a novel approach to genomic analysis using tools from the theory of thermodynamics. They adapt tools from thermodynamic formalism, particularly they adapt a well-known concept of ergodic theory and call it topological pressure. This feature is given by a weighted measure of complexity of a finite sequence. Two goals of the study were achieved: (1) using the topological pressure, applied on the human genome, it is predicted the coding sequence density on other genomes. This result gives a key practical advantage to the present approach, because it is possible to use only a single moderately phylogenetically distant informant genome as training data; (2) using the theory of thermodynamic formalism, the data encoded on the basis of topological pressure are turned into a probability distribution, which measures the coding potential of sequences of nucleotides of length between 750 and 5,000 bp. Another advantage of the present approach is its speed in calculation: the topological pressure can predict a coding sequence density for a genome in a matter of seconds, while ab initio prediction programs take a few hours, and evidence-based methods can take weeks.