Detecting Alien Genes in Bacterial Genomesa

Abstract
We present new methods for calculating codon bias of a group of genes or an individual gene relative to a standard gene class. This method is suitable for identifying alien (e.g., horizontally transferred) and highly expressed genes. In yeast and several bacterial genomes, highly expressed genes typically include ribosomal protein genes, elongation factors, chaperonins (heat shock proteins), and a subset of genes involved in glycolysis generally essential in exponential growth. Highly expressed genes of the Synechocystis genome feature several photosystem II genes, and highly expressed genes in several methanogens (Methanococcus jannaschii, M. thermoautotrophicum) are essential for methanogenesis. Alien genes mostly consist of ORFs of unknown function, transposases, prophage genes, and restriction/modification enzymes. Notably, nuclear ribosomal proteins of yeast are highly expressed, whereas mitochondrial ribosomal protein genes appear to be alien genes. Alien genes often occur in clusters, suggesting in these cases that transfer events entail several genes.