Table 1

Comparison of conserved optimal codon numbers at paired and unpaired sites.
Method
 Algorithm
 ALL
 GC-ending
 AT-ending
 GC leu & lys
Genes shorter than 800 bp
RNAfold S. cerevisiae
 MFE
 0.646 ***
 0.624 **
 0.503 ***
 1.353 ***

Mc
 0.542 ***
 0.567 ***
 0.422 ***
 0.910 ***
RNAfold S. paravensis
 MFE
 0.667 ***
 0.608 ***
 0.542 NS
 1.137 ***

Mc
 0.560 ***
 0.571 ***
 0.407 **
 1.119 ***
RNAfold S. mikitae
 MFE
 0.653 ***
 0.590 ***
 0.544 *
 1.172 ***

Mc
 0.572 ***
 0.623 ***
 0.411 ***
 1.131 ***
RNAfold S. bayanus
 MFE
 0.640 ***
 0.597 **
 0.502 ***
 1.181 ***

Mc
 0.537 **
 0.557 ***
 0.401 ***
 1.003 ***
ALIfold
 MFE
 0.638 ***
 0.577 ***
 0.465 NS
 1.499 ***

Mc
 0.468 **
 0.444 ***
 0.326 ***
 0.997 ***
All genes
RNAfold S. cerevisiae
 MFE
 0.920 ***
 0.863 ***
 0.751 ***
 1.584 ***

Mc
 0.841 ***
 0.866 ***
 0.676 ***
 1.436 ***
RNAfold S. paravensis
 MFE
 0.878 ***
 0.826 ***
 0.733 ***
 1.497 ***

Mc
 0.819 ***
 0.822 ***
 0.659 ***
 1.460 ***
RNAfold S. mikitae
 MFE
 0.912 ***
 0.814 ***
 0.790 NS
 1.160 ***

Mc
 0.839 ***
 0.869 ***
 0.740 NS
 1.404 ***
RNAfold S. bayanus
 MFE
 0.904 ***
 0.855 ***
 0.742 ***
 1.590 ***

Mc
 0.833 ***
 0.840 ***
 0.675 ***
 1.455 ***
ALIfold
 MFE
 0.899 ***
 0.759 **
 0.739 NS
 1.937 ***

Mc
 0.770 ***
 0.724 ***
 0.645 ***
 1.529 ***

I combine contingency tables for all amino acids and genes (ALL) and subsets of amino acids with GC- and AT- ending optimal codons (leu and lys are treated separately, as these two GC-ending amino acids behave very opposing, see below Table 2). Mantel Haenzsel estimators and significances are presented, WMH <1 = lower optimal codon use at paired than at unpaired sites.

* < 0.05, ** < 0.01, *** < 0.005, NS = not significant. Structure prediction is based on ALIfold and RNAfold using MFE and McCaskill's (Mc) algorithm.

Stoletzki BMC Evolutionary Biology 2008 8:224   doi:10.1186/1471-2148-8-224