Fig 1.
Classification of the genetic code according to codon GC content.
The G and C nucleotides are highlighted in red. The GC contents of codons are expressed as a percentage in the last column. The 20 standard amino acids were classified according to the number of G or C nucleotides divided by the number of nucleotides in their respective codons.
Fig 2.
Distribution of the genomic GC content of 192 studied prokaryotes.
The boxplots show the results of 12 phyla with at least three organisms with respect to their genomic GC content. The first boxplot corresponds to the genomic GC content values of all the studied microorganisms, which range from 20 to 74%. The boxplots were ordered in accordance with the GC content values of the first quartile.
Fig 3.
Genomic GC content bias effects on three amino acid frequencies for proteome and secondary structures of proteins.
The relative frequencies of amino acids in proteomes (a, c, and e) or in the secondary structures of proteins (b, d, and f) of 192 prokaryotes are shown. The amino acids of the figures were chosen considering their tendencies to be part of each of the secondary structures of proteins (b, d, and f, for alpha-helix strong former, indifferent, or beta-sheet strong former, respectively) or considering the GC content of the codons by which they are encoded (a, c, and e, for high, neutral, and low GC content codons, respectively). As a reference, the regression lines of the amino acid frequencies obtained in the proteome analyses are shown in purple in figures b, d, and f. The conformational parameters of amino acids described by Chou and Fasman [25, 27] are shown at the top of each graphic.
Fig 4.
Effect of the genomic GC content on the amino acid conformational parameter values.
The CP values for Gly, Pro, Tyr, Ile, and Ser in alpha-helices (red circles), beta-sheets (green circles), and random coils (blue circles) are evaluated as a function of the genomic GC content. The regression lines of the points are represented by solid lines, while the CP values reported by Chou and Fasman [25, 27] are represented by hashed lines.
Fig 5.
Relative frequencies of the secondary structure of proteomes as a function of the genomic GC content.
The relative frequencies of alpha-helix, beta-sheet, and random coil secondary structures in our studied organisms were evaluated and plotted against their corresponding genomic GC content. A linear regression was obtained for each of the secondary structure elements. The slope value (m) for each linear regression is indicated in the figure and the linear regression with a p-value <0.001 is found in S7 Table.
Fig 6.
Comparative linear regression between COGs without and with bias imposed by the GC content of their respective genes.
a) COG0002: Acetylglutamate semialdehyde dehydrogenase, and b) COG3228: Uncharacterized protein conserved in bacteria, respectively. The regression lines for each secondary structure are shown as solid lines. A complete set of figures is available from our GCto2D web server at https://biocomputo.ibt.unam.mx/gcto2d/.