Reasoning from the interface
Rational design replaces search with reasoning. Given a structure of the antibody–antigen complex, the scientist can see which antibody residues contact the antigen and propose specific changes — a new hydrogen bond, a filled cavity, removal of a repulsive charge. The approach is powerful precisely because it is targeted, and fragile precisely because it inherits every error in the structure it reasons from.
The maturation loop
Affinity maturation is a cycle of mutation, expression, measurement, and selection. Each round explores the neighborhood of the current sequence rather than the whole sequence space, which is what makes the search tractable. The measurement step is not optional: without it, there is no way to know whether a proposed change helped or hurt.
Two ways of choosing the next variant
Rational, structure-guided
- Starts from a known or modeled interface structure
- Changes residues chosen for a specific structural reason
- Few variants tested per round, each with a rationale
- Fails when the structure is wrong or the energetic effect is misjudged
Affinity maturation
- Starts from a candidate that already binds
- Introduces mutations without requiring a structural reason for each
- Many variants tested per round, best ones carried forward
- Improves gradually; each round explores only a local neighborhood
Neither approach removes the scientist from the loop. Someone still decides which residues to target, how many mutations to introduce at once, how stringent the selection should be, and when the result is good enough to stop.