How does affinity maturation work, and can design replace the campaign?
Affinity maturation improves how tightly a molecule binds its target. In an animal it happens by somatic hypermutation and selection; in a laboratory it is usually reproduced by display, building a library of variants and panning it. We do it by design instead: we model the interface, propose the substitutions that should strengthen it, rank them, and test a small panel rather than a library.
How it actually works
In nature, B cells mutate their antibody genes and the variants that bind better are selected to proliferate. Display reproduces that loop in vitro, which is why it works and also why it is slow: you are searching blind, so you need a very large library and several rounds.
Design changes what you search. The interface is a finite set of contacts, and most of the affinity sits in a handful of them. Modelling tells you which contacts are weak, which positions could carry a better side chain, and which substitutions would cost you elsewhere. You end up testing tens of candidates instead of millions.
Design does not replace measurement. It replaces the blind part of the search.
Affinity is not the only axis, and that matters here
Pushing affinity is the easiest way to break a molecule. A tighter interface is often a more hydrophobic one, and a more hydrophobic interface aggregates. Matured candidates that lose expression or gain a liability are a classic and expensive outcome.
We profile developability on every matured variant, not at the end but in the ranking itself, so the panel you receive has already been filtered for the molecules you could actually develop.
Where maturation is the wrong answer
If the binding site is in the wrong place, no amount of maturation will fix it. A molecule that binds the wrong epitope, or an epitope that is occluded in the assembled target, has a geometry problem, not an affinity problem. We say so at the assessment stage rather than selling you a campaign that cannot succeed. In that case the honest answer is to design a new binder against the epitope you actually need.
What you receive
- A ranked panel of matured variants, not a single bet. We generate and filter on the order of five hundred candidates to get there.
- The interface analysis: which contacts carry the binding, which positions were changed and why.
- The measured comparison against the parent: affinity, expression, thermal stability, aggregation behaviour.
- The developability profile of each candidate, so a gain in affinity is not a hidden loss elsewhere.
- Full ownership of the sequences, the data and the resulting intellectual property.
How it works
- Assessment: We read the molecule and its interface, and tell you whether maturation is the right lever at all. Free.
- Design: Substitutions proposed on the interface, ranked under a developability constraint.
- Measurement: A small panel expressed and compared with the parent at the bench.
Why clients come to us
We search a small space on purpose. Modelling the interface is what lets you test tens of candidates instead of a library.
We will tell you when maturation is not your problem. A geometry problem does not respond to affinity engineering, and finding that out after a campaign is the expensive way to learn it.
Developability is in the ranking, not in the post-mortem. A matured molecule you cannot express is not a result.
You own everything. Sequences, data and intellectual property are yours.
Send us your molecule and its target. We will read the interface and tell you whether maturation is the right lever at all, at no cost.