Codon frequency

Coding sequences have characteristic features such as codon frequencies, start- and stop-codons, upstream TATA-boxes, and splice-sites, which can be used to annotate an ORF (14). While each piece of evidence is not, by itself, conclusive, the combination of factors can lead to firm identification of genes. Unfortunately this method is prone to a large number of false-positives, with at least 75% of predicted genes typically being artefacts (15). 

The use of synthetic trinucleotides as building blocks would obviate the problems of stop codons or other unwanted amino acids turning up randomly in hypervariable sequences, and would indeed allow the predetermined definition of amino acid codon frequency bias. This has been demonstrated by Glaser et al. [97] for phage-display library construction see also Ref. 98 for a review of current DNA synthesis technology. 

ANACONDA then calculates the value of the Pearson s chi-squared statistic and the adjusted Pearson residual values. Pearson s statistic represents a global measure of the difference between observed and expected codon frequencies (20). 

One possible explanation is the simple observation that the G-rich codons do not code for the same amino acids as C-rich codons, but modelling based on codon frequencies shows that the predicted discrimination from this alone is not sufficient to account for the observation. An alternative explanation is based on the observation that the G-patterns would lead to potential quadruplexes in the mRNA strand in addition to the DNA duplex, whereas C-patterns could only lead to quadruplexes in the (complementary strand of) DNA. This evidence is consistent with an evolutionary pressure to reduce the number of quadruplexes allowed to form in mRNA. This may be particularly strong as it seems that RNA quadruplexes are more stable, both thermodynamically and kinetically, than their DNA counterparts. To date, there has been relatively little work focused on RNA quadruplexes, although it has evoked some interest. However these results suggest that RNA quadruplexes could play a significant physiological role, and should be investigated further. 

Figure 11.4. Histograms comparing average codon frequencies of orthologous genes from human and calf (Cruveiller and Bernardi, unpublished results).

Figure 11.5. Hii>togr

Figure 11.9. A. Codon frequencies of human (top) and Xenopus (middle) GC-poor (0-60% GC3) orthologous genes. The bottom histogram shows the differences in codon frequencies. B. Codon frequencies of human (top) and Xenopus (middle) GC-rich (60-100°/i GC3) orthologous genes. The bottom histogram shows the differences in codon frequencies. (From Cruveiller et al., 2000).

Cruveiller S., D Onofrio G., Bernard G. (2000). The compositional transition between the genomes of cold- and warm-blooded vertebrates codon frequencies in orthologous genes. Gene 261 71-83. 

The previously discussed formulations used to delineate reading frames can be used In conjunction with the trlmer preferences. The previous codon frequencies are replaced by the probability that a third amino acid follows the two previous. For example, a particular reading frame of a nucleotide sequence yields AVLCDXFALL where X Is a STOP codon. Then, the fcl(l) In the Ml(l) formula (given previously) Is replaced by (F Y l) where F vl observed frequency of AVL In the... 

In order to optimize codon usage in a synthetic gene, the frequency of codon usage must be included with the run (Fig. 3). Codon frequency is determined using the method of Nakamura et al. (8). The codon frequency table also directs the reverse translation of amino acids to codons. The format is that of the GCG Wisconsin Package. 

DNAWorks contains a small set of preloaded codon frequency tables for common organisms. A user can also manually type in codon frequencies or upload a file (Fig. 3). Nucleotide-only sequences do not require a codon frequency table. 

TmPrime includes a codon optimizing feature. It implements global codon optimization that replaces each codon based on the organism-specific codon frequencies using the organism-specific codon data in the Codon Usage Database (http//www.kazusa. 

The codon frequency is given as fractions of codons for a given amino acid. Data are for pre-pro-proteins. Data for human C3 are from (17). Data for cobra C3 are from (25). 

---