The surprising negative correlation of gene length and optimal codon use - disentangling translational selection from GC-biased gene conversion in yeast

Table 4 Controlling the effect of gene position.

	Gene length	Expression	Recombination
F _OP AT N = 147193 codons R² = 0.260* P(GL)NS,P(Expr), P(dmc1, mre11₀)**	-0.0480NS	+0.3696***	Co:-0.1064* nCo:-0.0922* spo11:-0.0442NS dmc1:-0.1224*** mre11₀:-0.0520* mre11₆:-0.0922***
F _OP GC N = 180789 codons R² = 0.445* P(GL)NS, P(Expr), P(mre11₀), P(dmc1)***	-0.1289***	+0.3839***	Co:+3284* nCo:+0.2472* spo11:+0.3909* dmc1:+0.4829* mre11₀:+2648* mre11₆:+0.2669*
F _GC Opt N = 70401 codons R² = 0.174* P(GL)NS, P(Expr)NS P(dmc1, mre11₀)*	-0.0751**	+0.1216***	Co:+0.2735* nCo:+0.1843* spo11:+0.3232* dmc1:+0.3837* mre11₀:+0.2104* mre11₆:+0.2368*
F _GC Non-Opt N = 126878 codons R² = 0.224* P(GL)NS, P(Expr)* P(dmc1, mre11₀)***	-0.0481NS	+0.0410NS	Co:+0.2990* nCo:+0.2339* spo11:+0.3506* dmc1:+0.4390* mre11₀:+0.2311* mre11₆:+0.2754*

Spearman Rank Correlations between frequencies of optimal and GC-ending codons with gene length, expression, and various recombination measures after controlling for a potential effect of gene position. The same 1571 genes were used for all measures. Presented are also MR results of variables whose log-transformation did not grossly deviate from a normal distribution. *P < 0.05, **P < 0.01,***P < 0.001

ISSN: 2730-7182