Signal and Noise in Gene Length Distributions of Decoded Genomes.

摘要

Development of objective measures for comparing genome data has become an important goal in computational genomics. Gene length distribution histograms derived from the predicted protein-coding gene populations of decoded genomes were considered as tools for genomic comparison. The first and second moments of the histograms map the decoded genomes as isolated points onto an Euclidean plane. The clustering of the genomic points on the plane preserves the separation of the three domains of life. The mapping of genomes onto a plane allows for the use of vector analysis in the study of genomic evolution.