Mutation Rate Tracks Predictability

01The evidence

• Populations facing environmental change take multiple evolutionary routes to evolvability, and the route depends on the structure of the change rather than only on its magnitude.PNAS, 2024
• Adaptation to different rates of environmental change has been demonstrated directly: algal populations exposed to fast and slow fluctuations diverge in their adaptive strategies within a few hundred generations.Evolution, 2009

02The tension

Mutation rates vary by three orders of magnitude across microbes, and the standard explanations are about linkage: a mutator arises, hitches a ride with a beneficial allele, and spreads. That explains how a mutator can spread but not why some lineages maintain high rates for millions of years while others refine proofreading to the physical limit.

03The theory

Treat mutation as exploration and replication fidelity as a cost. Lower fidelity buys faster exploration and pays a load in deleterious mutations. The value of exploration depends on how long a good genotype stays good, which is exactly the autocorrelation time of the environment.

If the optimum moves on a timescale much longer than the time to fix a beneficial mutation, exploration is wasted and selection drives fidelity up. If the optimum moves faster than fixation, a lineage that cannot explore is outrun, and high mutation is favoured. The crossover produces an inverse relation between mutation rate and correlation time.

This reframes mutators as an adaptive allocation rather than an accident. It also predicts hysteresis: lineages that have been in stable environments for a long time should be slow to raise mutation rate when the environment destabilises, because the machinery for fidelity has been optimised and the regulatory paths to reduce it may have decayed.

The test is comparative and experimental at once. Comparative: measure niche autocorrelation from long-term environmental records and regress mutation rate against it across taxa. Experimental: evolve the same ancestor under predictable and unpredictable fluctuations at matched variance, and see whether mutation rate diverges.

04What this is not

05Predictions

• P1Across fifty microbial species with instrumented environments, mutation rate should correlate inversely with niche autocorrelation time after controlling for genome size and GC content.
• P2Experimental evolution under predictable versus unpredictable fluctuations at matched variance should produce divergent mutation rates within two thousand generations.
• P3Pathogens that move from variable environmental reservoirs into predictable hosts should lose mutator alleles on a measurable timescale.

06What would refute it

No residual correlation once genome size and phylogeny are controlled would leave mutation rate explained by linkage and drift alone.

07What it would change

It would let us forecast resistance evolution from environmental data, and reframe proofreading as a dial that evolution sets rather than a fixed species trait.