Why does the tightest binder often fail to neutralise?
The problem with affinity-led screening
Immunodominance is a real force. One epitope tends to dominate the response, so a panel ranked on binding fills up with variations on the same clone. If that epitope happens to be functional, the campaign looks brilliant. If it does not, the campaign produces a lot of excellent antibodies that do nothing, and the failure is only discovered in the functional assay at the end.
The evidence that affinity is a poor proxy for function is not just anecdotal. In our own published work on methylation-specific antibodies, the tightest binder in the panel — 0.90 nM — was nearly unusable in the application, while 14 nM and 54 nM clones worked cleanly. The structures explain why.
How we do it
- Recover breadth, then organise it. We keep a broad, diverse panel and cluster it by HCDR3 similarity. Clones sharing an HCDR3 lineage tend to share an epitope, so lineages are a practical proxy for functional families — and they tell you how many genuinely different things you have, not just how many wells were positive.
- Test lineages, not clones. We work with you to test representatives from each lineage and home in on the ones that drive the function you need. That is a much smaller experiment than screening everything.
- Use the non-functional clones. They are not waste. Clones against the dominant epitope can be used to mask it during a further selection round, forcing recovery of antibodies against novel epitopes that were previously outcompeted.
- Select for interference directly. For ligand-receptor blockers we use the receptor or the ligand itself as a reagent during panning, so clones that disrupt the interaction are enriched by the selection rather than discovered afterwards.
- A lineage-clustered panel with sequences, not a single clone
- Binding data across the panel
- Epitope-binning information implied by the lineage structure
- Recombinant material for the clones you advance
Need antibodies that do something, not just bind?
Tell us the target, the functional endpoint, and the assay you will judge it in. A PhD scientist reads it and replies within two business days. No NDA needed to start.
Talk to a scientist →Frequently asked questions
Why do you deliver a panel rather than the single best binder?
Because affinity does not predict function. The clone with the tightest KD is frequently not the one that neutralises, because neutralisation depends on where it binds, not how hard. Clustering by HCDR3 lineage surfaces functionally distinct families so the functional screen has something to choose between.
What if all the clones hit the same dominant epitope?
That is a common failure and it has a fix. Non-functional clones can be used to mask the dominant epitope during a further round of selection, forcing recovery of antibodies against the epitopes that were being outcompeted.
Can you select directly for blocking?
Yes. For ligand–receptor pairs we use the receptor or the ligand itself as a reagent during selection, so clones that interfere with the interaction are enriched rather than merely hoped for.
Who runs the functional assay?
Usually the sponsor, because it is your biology and your assay. We work with you to decide which lineages to test and in what order. If the assay is straightforward we can sometimes run it, which is worth discussing at scoping.
Can you improve a clone that blocks weakly?
Yes. Affinity maturation and kinetic tuning are the usual next steps, and iterative re-maturation is how we drove broad neutralisation across a moving target set in our published SARS-CoV-2 work.