DAMAGE TO PROTEINS resulting from gene mutations or mistranslations is minimized when the errors substitute amino acids with similar affinities for water (hydrophobicity). If we define a code's error value as the average change in amino acid hydrophobicity caused by all possible single-base changes within all codons of a code, then a high error value indicates that a code is very vulnerable to errors, and a low value means that a code minimizes their harm. We generated a large random sample of possible codes and found that only 100 of one million alternatives had a lower error value than nature's code (top). When we factor in real-world patterns in the way genes mutate and are mistranslated, nature's code outperforms all but one in a million of the alternatives (bottom).

Comments