Codon Preference is but an Illusion Created by the Construction Principle of Coding Sequences

Modern coding sequences are in the periodicto-chaotic transition. In the case of two related sequences for lens alpha A-crystallin and small heat shock protein, the original repeating units were heptameric in length. Accordingly, base trimers that were parts of heptameric units recurred far more frequently than those that were not included. In the crystallin coding sequence, CTG trimer recurred 21 times, and TCT and TCC trimers recurred 17 times each. By contrast, CTA and TCG, although related to the above, recurred only 4 and 3 times, respectively. It is a small wonder that 10 of the 16 leucine residues were encoded by CTG, whereas none was encoded by CTA, and that 17 of the 23 serine residues were encoded either by TCT or by TCC, whereas only 1 was encoded by TCG. In the small heat shock protein coding sequence, however, AGC became parts of the two prominent heptameric recurring units. Not surprisingly, 10 of the 22 serine residues were now encoded by AGC. In conclusion, the so-called codon preference is a mere reflection of the construction principle of coding sequences and has very little to do with selection per se.