Analysis of a Circular Code Model

A circular code has been identified in the protein (coding) genes of both eukaryotes and prokaryotes by using a statistical method called trinucleotide frequency (TF) method [Arquès & Michel (1996). J. theor. Biol.182, 45–58]. Recently, a probabilistic model based on the nucleotide frequencies with a hypothesis of absence of correlation between successive bases on a DNA strand, has been proposed by Koch & Lehmann [(1997). J. theor. Biol.189, 171–174] for constructing some particular circular codes. Their interesting method which we call here nucleotide frequency (NF) method, reveals several limits for constructing the circular code observed with protein genes.