Synonymous substitutions (silent, do not change amino acids) accumulate faster than non-synonymous substitutions due to weaker purifying selection. The ratio dN/dS (non-synonymous to synonymous rate) reveals selection pressure: dN/dS < 1 indicates purifying selection, dN/dS > 1 suggests positive selection for protein change.
From the genetic code, you know that most amino acids are encoded by multiple codons — for example, leucine has six codons (UUA, UUG, CUU, CUC, CUA, CUG). This redundancy means that some nucleotide changes in a protein-coding gene alter the amino acid sequence while others do not. A synonymous substitution (also called a silent substitution) changes a codon to another codon that specifies the same amino acid — for instance, CUU → CUC both encode leucine. A nonsynonymous substitution changes the amino acid — for instance, CUU (leucine) → CCU (proline). This distinction, rooted in the structure of the genetic code, turns out to be one of the most powerful tools in molecular evolution.
The logic is straightforward: synonymous changes leave the protein untouched, so they are largely invisible to natural selection and accumulate at a rate close to the neutral mutation rate. Nonsynonymous changes alter the protein, and most such alterations are harmful, so purifying selection removes them. The result is that the nonsynonymous substitution rate (dN) is typically much lower than the synonymous substitution rate (dS). When you compare the same gene across two species, counting synonymous and nonsynonymous differences separately gives you a direct window into the selective forces acting on that protein.
The ratio dN/dS (also called ω or Ka/Ks) is the key metric. When dN/dS < 1, nonsynonymous changes are being removed faster than they accumulate — this is the signature of purifying selection constraining the protein. The lower the ratio, the stronger the constraint. When dN/dS ≈ 1, nonsynonymous and synonymous changes accumulate at the same rate, suggesting the protein (or that region of it) is evolving neutrally — amino acid changes have no fitness effect. When dN/dS > 1, nonsynonymous changes are accumulating *faster* than synonymous ones, which can only happen if natural selection is actively favoring amino acid changes — the hallmark of positive selection driving protein adaptation.
To make this concrete, consider a comparison between humans and mice. Histone genes, which encode proteins critical to chromosome structure, have dN/dS ratios near 0.01 — almost all amino acid changes are lethal and removed. Olfactory receptor genes have dN/dS around 0.3–0.5 — moderately constrained but tolerating some change. Genes involved in immune defense or reproduction sometimes show dN/dS > 1 in specific regions, indicating an evolutionary arms race where protein change is actively favored. By computing dN/dS across thousands of genes, researchers can identify which proteins are under the strongest constraint, which are evolving neutrally, and which are sites of adaptive evolution — all from sequence data alone.