Claim № 6.02  · Pillar 6 · Case file

“Natural selection removes harmful mutations as fast as they arise, so genomes do not degrade over time.”

Disciplines: Genetics · Population genetics
Status: reviewed 2026-11-15
UnderExaminationTwo claims · two grades

§1 The claim, as taught

Natural selection acts to promote traits and behaviors that increase an organism’s chances of survival and reproduction, while eliminating those traits and behaviors that are detrimental to the organism. However, natural selection can only, as its name implies, select—it cannot create.Clark MA, Douglas M, Choi J. Biology 2e. OpenStax (2018), ch. 19 Introduction, “The Evolution of Populations.”

§2 The strongest evidence offered

This debate is unusual: the problem was posed by mainstream geneticists, and the critics' case uses their equations. The Lab's genetic entropy simulator lets you test the claims yourself.

Observed

Dozens of new mutations each

Each person carries roughly 60–70 new mutations not present in either parent. How many affect fitness depends on how much of the genome is functional, which is disputed (see Pillar 8).
Kong et al., Nature 488 (2012); Keightley, Genetics 190 (2012).
Observed

Most effects are slight

Of the mutations that change fitness, most are harmful and most of those are only slightly harmful. Mutations with very slight effects are nearly invisible to selection and can drift to high frequency, especially in small populations.
Ohta, Nature 246 (1973); Eyre-Walker & Keightley, Nature Reviews Genetics 8 (2007).
Observed

Decline without selection

When selection is nearly switched off, by breeding lines through single individuals, fitness falls steadily in many organisms. In asexual lineages, harmful mutations can ratchet up irreversibly.
Halligan & Keightley, Annu. Rev. Ecol. Evol. Syst. 40 (2009); Chao, Nature 348 (1990).
Inferred

A paradox named by the mainstream

Alexey Kondrashov asked why, given the number of very slightly harmful mutations, we have not “died 100 times over.” He called the paradox real and proposed soft selection and interacting mutations as ways out. Michael Lynch has warned that modern humans may be accumulating load measurably, because medicine relaxes selection.
Kondrashov, Journal of Theoretical Biology 175 (1995); Lynch, Genetics 202 (2016).
Inferred

Proposed resolutions

Mainstream models show that load can be tolerated if selection acts on relative fitness (soft selection), if mutations interact so selection removes several at once, or if beneficial and compensatory mutations keep pace. Degraded worm lines recovered full fitness within 80 generations in large populations.
Lesecque, Keightley & Eyre-Walker, Genetics 191 (2012); Estes & Lynch, Evolution 57 (2003).
Extrapolated

Genetic entropy

John Sanford argues that the human genome is degenerating, and that this is incompatible with millions of years of evolution. His case rests on simulations (Mendel's Accountant) and on an analysis of the H1N1 flu virus. Most population geneticists dispute his premises, especially how much of the genome is functional.
Sanford, Genetic Entropy (2005, 2014); Carter & Sanford, Theoretical Biology and Medical Modelling 9 (2012).

§3 The grade: two claims hiding in one

Strongly harmful and very slightly harmful mutations behave differently. Graded separately: Hover or tap a rung to see why it got its score. How the grading works →

Selection efficiently removes strongly harmful mutations

BStrongly supported
Observation
///
Observed: severe genetic diseases stay rare.
Hypothesis
///
Precise mutation–selection balance models.
Prediction
//
Predicted equilibrium frequencies roughly match observed disease frequencies.
Experiment
//
Lab populations and mutation-accumulation lines confirm it.
Replication
///
Replicated across many organisms.
Falsifiability
//
Clear predictions that could fail.
Method score 15/18

Selection removes slightly harmful mutations as fast as they arise, so genomes do not degrade

DExtrapolation
Observation
/
Not observed in long-lived, slow-breeding species; long-term fitness is rarely measured.
Hypothesis
//
Specific mechanisms are proposed: soft selection, epistasis, compensation.
Prediction
/
The theory's own nearly-neutral models predict that slight harms accumulate in small populations.
Experiment
/
Recovery experiments exist, but in large populations over short times.
Replication
/
Load estimates differ widely depending on assumptions.
Falsifiability
/
A new mechanism is proposed whenever the calculated load is too high.
Method score 7/18
ADemonstratedBStrongly supportedCInferenceDExtrapolationEAssumption

§4 What it would take

To show that genomes do not degrade, the following would be needed:

  • Measure the slight effectsDirect estimates of how many new mutations have effects too small for selection to see, from very large samples.
  • Long-term fitness dataMultigenerational measurements of fitness in a slow-breeding sexual species, not only growth rates in microbes.
  • Show the resolution worksEvidence that soft selection or synergistic epistasis operates in real populations at the strength the models need.

§5 Another reading of the same data

The textbook picture is tidy: harmful mutations are weeded out, helpful ones kept. Population geneticists have long known that it is not so simple. Hermann Muller, who won a Nobel Prize for showing that radiation causes mutations, wrote in 1950 about “our load of mutations” and worried about its growth. The trouble is the very slightly harmful ones. They are too weak for selection to see individually, but there are many of them, and they add up.

The mainstream response is serious and partly persuasive. If selection mostly ranks individuals against one another (soft selection), a population can carry a large load without shrinking. If harmful mutations interact, selection can remove several at once. Beneficial and compensatory mutations push the other way. Each mechanism is real. What has not been shown is that they operate strongly enough, in a slow-breeding species like ours, to hold the line indefinitely.

Read through a recent-creation lens, genomes began in very good condition and have been running down since: slowly, not catastrophically, through the slow accumulation of slight damage. That would fit the declining lifespans recorded in Genesis, and it would make the human genome's current condition a few thousand years of wear rather than millions of years of balance. The reading has its own burden. It needs the share of functional DNA to be high, which Pillar 8 examines, and it needs better measurements of the slight effects than anyone yet has.

§6 Reasoning check

The grade weighs the evidence. This check looks at the reasoning: the step from the evidence to the conclusion. The same tests are applied to both readings. How the reasoning check works →

In the claim as taught
  • Possible

    Soft selection and synergistic epistasis were proposed largely because the calculated load was too high. Both are real phenomena, which keeps this at possible, but how strongly they act in nature has not been measured.

  • Possible

    Textbooks report that selection removes harmful mutations, and rarely mention that the theory's own models predict accumulation of the slightly harmful ones.

  • Possible
    In the evidence: Genetic entropy

    From simulation results under chosen parameters, and one viral lineage, to the trajectory of the whole human genome. The parameters that decide the answer, such as the share of functional DNA, are the ones in dispute.

In the other reading
  • Possible

    Genetic entropy predicts decline, and the H1N1 lineage declined in virulence. But rising immunity and better medicine predict the same, so the observation alone does not decide between them.

§7 Questions to ponder

  • If the founders of population genetics worried about mutational load, why do textbooks rarely mention it?
  • Is a population that keeps its numbers while accumulating slight damage stable, or degrading slowly?
  • What would we expect to see if life were recent and genomes started out in very good condition?

§8 Sources

Supporting the mainstream claim
  1. Muller HJ. “Our load of mutations.” American Journal of Human Genetics 2:111–176 (1950).C
  2. Ohta T. “Slightly deleterious mutant substitutions in evolution.” Nature 246:96–98 (1973).B
  3. Kondrashov AS. “Contamination of the genome by very slightly deleterious mutations: why have we not died 100 times over?” Journal of Theoretical Biology 175:583–594 (1995).C
  4. Eyre-Walker A, Keightley PD. “The distribution of fitness effects of new mutations.” Nature Reviews Genetics 8:610–618 (2007).B
  5. Kong A et al. “Rate of de novo mutations and the importance of father's age to disease risk.” Nature 488:471–475 (2012).B
  6. Lesecque Y, Keightley PD, Eyre-Walker A. “A resolution of the mutation load paradox in humans.” Genetics 191:1321–1330 (2012).B
  7. Estes S, Lynch M. “Rapid fitness recovery in mutationally degraded lines of Caenorhabditis elegans.” Evolution 57:1022–1030 (2003).B
  8. Lynch M. “Mutation and human exceptionalism: our future genetic load.” Genetics 202:869–875 (2016).C
Critical & alternative perspectives
  1. Carter RW, Sanford JC. “A new look at an old virus: patterns of mutation accumulation in the human H1N1 influenza virus since 1918.” Theoretical Biology and Medical Modelling 9:42 (2012).C
  2. Basener WF, Sanford JC. “The fundamental theorem of natural selection with mutations.” Journal of Mathematical Biology 76:1589–1622 (2018).C
  3. Sanford JC. Genetic Entropy & the Mystery of the Genome. Ivan Press (2005); later editions titled Genetic Entropy.D

Source grades are provisional. In production each links to a full published evaluation that readers can challenge.

§9 Your notes