On 2015 Aug 26, Donald Forsdyke commented:

DIMINISHED ROLE FOR CONVENTIONAL NATURAL SELECTION

This study decisively demonstrates both that nucleic acid level forces drive amino acid composition, rather than the converse, and that, in this respect, higher oligonucleotide frequencies are more powerful than mononucleotide frequencies (base composition). This is consistent with the case, made on different grounds, that oligonucleotide frequencies drive mononucleotide frequencies (summarized in Forsdyke 2011). Furthermore, there is now better support for Grantham’s “genome hypothesis” that natural selection by way of the conventional environment, may be secondary to some other form of selection that relates to speciation (see comment on Goncearenco A, 2014). Indeed, in some cases, amino acids in a protein may be mere “place holders” - there to serve the needs of the genome (Rayment JH, 2005).

Of course, some adaptation takes place at the protein level, but that the authors’ reading frame-specific analysis provides “contravening evidence” against the power of oligonucleotides is not readily apparent. In thermophiles the low frequency of TpA overlapping successive codons (e.g. NNT,ANN, …), and the depletion of ApT when positioned within a codon (e.g. NAT, NNN, … ), are easily explained by the pressure on thermophiles to purine-load their coding sequences (i.e. there is a nucleic acid level selective pressure). Thermophiles can best achieve this, without imposing excessively on amino acid composition, by incorporating purines in third codon positions. Thus, instead of the classical distribution of purines (R) and pyrimidines (Y) in codons (e.g. RNY, RNY, …), thermophiles tend to follow the RNR rule (e.g. RNR, RNR, … ) (see Lambros RJ, 2003).

Forsdyke DR: Evolutionary Bioinformatics. 2nd edition. New York: Springer, 2011.

On 2015 Aug 26, Donald Forsdyke commented:

DIMINISHED ROLE FOR CONVENTIONAL NATURAL SELECTION

This study decisively demonstrates both that nucleic acid level forces drive amino acid composition, rather than the converse, and that, in this respect, higher oligonucleotide frequencies are more powerful than mononucleotide frequencies (base composition). This is consistent with the case, made on different grounds, that oligonucleotide frequencies drive mononucleotide frequencies (summarized in Forsdyke 2011). Furthermore, there is now better support for Grantham’s “genome hypothesis” that natural selection by way of the conventional environment, may be secondary to some other form of selection that relates to speciation (see comment on Goncearenco A, 2014). Indeed, in some cases, amino acids in a protein may be mere “place holders” - there to serve the needs of the genome (Rayment JH, 2005).

Of course, some adaptation takes place at the protein level, but that the authors’ reading frame-specific analysis provides “contravening evidence” against the power of oligonucleotides is not readily apparent. In thermophiles the low frequency of TpA overlapping successive codons (e.g. NNT,ANN, …), and the depletion of ApT when positioned within a codon (e.g. NAT, NNN, … ), are easily explained by the pressure on thermophiles to purine-load their coding sequences (i.e. there is a nucleic acid level selective pressure). Thermophiles can best achieve this, without imposing excessively on amino acid composition, by incorporating purines in third codon positions. Thus, instead of the classical distribution of purines (R) and pyrimidines (Y) in codons (e.g. RNY, RNY, …), thermophiles tend to follow the RNR rule (e.g. RNR, RNR, … ) (see Lambros RJ, 2003).

Forsdyke DR: Evolutionary Bioinformatics. 2nd edition. New York: Springer, 2011.