sequence homology

Two cousins may have the same surname not because they happen to look alike, but because they descend from the same grandparent. In biology, two sequences are called homologous when they share that kind of common ancestry — they are not similar by coincidence; they are two surviving versions of one ancestral sequence. Homology is a statement about history, not just about looking alike.

Here is the careful part that trips many people up. Similarity is what you measure — you align two sequences and count how many positions match. Homology is what you infer from that similarity — the conclusion that the two share a common ancestor. Two sequences are either homologous or not; there is no such thing as being 65 percent homologous (that figure is percent identity, a measure of similarity). When a long stretch of two genes matches far more than random chance would allow, the simplest explanation is shared descent, so we call them homologous. Homology can be partial: two proteins may share one homologous domain, like a borrowed module, while the rest of each protein has a separate origin.

Why it matters: homology is the bridge that lets discoveries travel between species. If a well-studied human gene is homologous to a yeast gene, the yeast version often does a related job, so experiments in yeast illuminate human biology. Recognizing homology by sequence search (using tools that score alignments) is the first thing a biologist does with a new sequence. The common error to avoid is calling things homologous when they merely resemble each other for functional reasons — an analogy, like wings on birds and insects, is convergent and not homologous.