Natural populations maintain equilibrium between mutation introducing deleterious alleles and selection removing them. At equilibrium, mutation frequency equals loss due to selection, creating stable allele frequencies that depend on mutation rate and selection strength.
You already know from population genetics that allele frequencies change through drift and selection, and that the selection coefficient (s) quantifies how much a deleterious allele reduces fitness. Now consider a paradox: if selection removes harmful alleles every generation, why do genetic diseases persist at all? The answer is that mutation keeps reintroducing them. Mutation-selection balance is the equilibrium where the rate of new deleterious alleles entering the population exactly matches the rate at which selection purges them.
The math is elegantly simple. For a recessive lethal allele, the equilibrium frequency (q̂) is approximately √(μ/s), where μ is the mutation rate per generation and s is the selection coefficient. For a dominant deleterious allele, q̂ ≈ μ/s. These formulas tell you two important things. First, even strong selection (large s) cannot drive a deleterious allele to zero as long as mutation keeps feeding it back in. Second, the weaker the selection against an allele (smaller s), the higher its equilibrium frequency — because selection removes it more slowly while mutation introduces it at the same rate.
Consider a concrete example: cystic fibrosis. The CF allele has a mutation rate of roughly 10⁻⁶ per generation and is effectively recessive lethal (s ≈ 1 for homozygotes in historical populations). The predicted carrier frequency is √(10⁻⁶/1) = 0.001, or about 1 in 1,000. The observed frequency is actually much higher (~1 in 25 in European populations), which signals that something beyond simple mutation-selection balance is at work — likely heterozygote advantage. This is exactly how the model is useful: deviations from the predicted equilibrium point you toward additional evolutionary forces.
The concept connects directly to what you will study next. When selection is very weak (s is tiny), drift in finite populations can overpower selection and allow mildly deleterious alleles to drift to unexpectedly high frequencies — the domain of slightly deleterious mutations and nearly neutral theory. Mutation-selection balance assumes selection is the dominant removal force, which works well in large populations. In small populations, that assumption breaks down, and the interplay between mutation, selection, and drift becomes the central story of molecular evolution.