# Avoiding Catastrophic Mutations Accurately Predicts Amino Acid to Codon Pairing

## TL;DR

This paper proposes a new hypothesis about how the genetic code evolved to avoid dangerous mutations involving stop signals and cysteine.

## Contribution

The Catastrophic Mutation Minimization Hypothesis explains codon assignments by minimizing mutation risks in early genetic code evolution.

## Key findings

- 16 of 20 amino acids in the genetic code align with the CMMH predictions.
- Two variations of the ancestral doublet code are more mutation-resistant than most random models.
- The ancestral code already had key modern features like start/stop signals and disulfide bonding.

## Abstract

DNA codon mutations involving Stop signals or the amino acid cysteine can be especially damaging. The former can break protein sequences or add extraneous amino acids. The latter can add or subtract disulfide bonds crucial in protein folding. We present a hypothetical scenario where Stop codons were present early in the evolution of the genetic code; and minimizing catastrophic mutations for code networks affected all subsequent amino acid/codon pairings. Predicted features of this “Catastrophic Mutation Minimization Hypothesis” (CMMH) are that: (1) Cysteine is mutationally adjacent to Stop, isolating a contiguous codon ‘neighborhood’ with high potential for catastrophe. (2) The sequence of amino acid additions order determines codon assignments through minimizing network-wide mutation costs. Overall, codon locations for 16 of the 20 amino acids in the genetic code are consistent with the CMMH, as are multiple other predictions. We propose an antecedent genetic code consisted of 16 doublet codons specifying 13–14 amino acids. Two variations of these networks are less susceptible to catastrophic mutations than 88.2–97.5% of randomly generated ones. Unlike some previous hypotheses, CMMH does not require the total replacement or rearrangement of amino acids at codons, with its disruptive potential for protein synthesis. Finally, the composition of this ancestral doublet genetic code has all the modern code’s utility: amino acids from four chemical types; start and stop signals; metal-binding ability; disulfide bridging for creating protein shapes; and possible epigenetic gene regulation. Thus, the modern code likely evolutionarily fine-tuned antecedent capabilities, rather than significantly increasing competence for making complex proteins.

The online version contains supplementary material available at 10.1007/s00239-025-10294-0.