conserved versus variable sites

Picture a recipe handed down through a hundred cooks. The line that says 'do not let it boil or it curdles' gets copied faithfully every time, because anyone who changes it ruins the dish. But the line about garnish — a sprig of parsley, or maybe basil — gets tweaked freely, because nothing depends on it. If you compared a hundred copies of the recipe, the warning line would be identical everywhere (conserved) and the garnish line would vary all over the place (variable). Genes work exactly the same way.

When you align the same gene or protein from many species, some positions are the same in every species and others differ from one to the next. A conserved site is a position that stays the same across the alignment; a variable site is one that changes. The reason is selection acting through survival. If a particular letter is essential — say it codes for an amino acid that holds an enzyme's active site in shape — then any mutation there tends to harm the organism, so those mutated lineages leave fewer descendants and the change is weeded out. The site stays frozen. Where the exact letter does not matter much, mutations accumulate harmlessly and the site drifts, becoming variable. So the pattern of conservation is a readout of which parts of a molecule matter.

This is one of the most useful ideas in molecular biology. Highly conserved regions point straight at the functional core of a protein — active sites, binding pockets, structural anchors — which is why aligning sequences across species is a standard way to find what a protein does and which mutations might cause disease. Conservation also drives the molecular clock and the detection of selection. One caution: a variable site is not necessarily unimportant, and a conserved site can occasionally be conserved by chance over short timescales — conclusions get stronger as you add more, deeply diverged species.