Your candidate, measured against the antibodies that made it to the clinic.
Binding data tells you a molecule works. It does not tell you it will survive development. Send us your candidates and we measure the liabilities that end programs later, in undiluted serum and plasma, label-free, on your own molecules. Every result comes back positioned against clinical-stage antibodies, so you learn where your molecule sits among drugs that reached patients.
6 Clinical-stage reference antibodies | 20 µg Per candidate and panel | 1 to 2 weeks From sample to data package | Full service Nothing to set up, nothing to maintain |
Your design loop stops one question too early
Discovery answers the binding question at scale. The question that decides whether a program survives is the one after it, and it is usually asked far too late.
1 Binding is no longer the bottleneck Sequence in, affinity out, hundreds of candidates per campaign. One published benchmark measured 1,320 designed binders against 16 targets. Not one of them was asked whether it is developable. | 2 The liabilities that kill programs are invisible in buffer Polyspecificity drives fast clearance and immunogenicity risk. Behavior in serum and plasma decides whether a clean buffer result survives contact with biology. Neither shows up in a target binding assay. |
3 Every liability is a separate late assay Polyspecificity, matrix behavior, neonatal Fc receptor (FcRn) binding and the Fc gamma receptor profile are four different plate assays on three different instruments, run weeks after affinity ranking on a short list that was already committed. | 4 So candidates are selected on incomplete information By the time the developability data arrives, the resources are spent. Focal molography measures these liabilities where they matter, in undiluted matrix, label-free, on the same chip and the same sample as the affinity readout. |
A number is not an answer. A position is.
A score on its own decides nothingA polyreactivity value of 0.34 tells a program manager nothing. The question that actually has to be answered is comparative: is this molecule stickier than antibodies that failed, or cleaner than antibodies that reached patients? That is a question about position, and it can only be answered against a reference set measured the same way, on the same surface, under the same conditions. A single number from a single assay, however precise, cannot answer it. | What your candidate is compared againstJain and colleagues surveyed 137 clinical-stage monoclonal antibodies in 2017 and showed that developability liabilities track with progression through clinical development. That distribution is the reference every candidate is reported against. We are extending it with a ladder of reference molecules measured on our own instrument, spanning the full range of what has been taken into humans, from among the stickiest antibodies ever brought into the clinic to Fc-engineered molecules with the longest half-lives. Your candidate is placed on that ladder, not just given a value. |
Four panels, one sample, one data package
Order a single panel or the full dossier. Each panel is documented in full, nothing here is a black box.
B1 Polyreactivity Non-specific binding across an 8-plex or 16-plex ligand panel, scored against the clinical-stage benchmark. The panel that catches fast clearance, aggregation and immunogenicity risk before they become someone else's problem. | B2 Plasma and serum behavior Bound versus free fraction in undiluted human matrix. Not a buffer proxy for physiology, but the measurement itself. |
C1 FcRn across species Neonatal Fc receptor binding for the species used in pharmacokinetic studies, so you learn whether your animal data will translate before you run it. | C2 Fc gamma receptor profile Engagement across the activating and inhibitory receptors that set the effector profile of an Fc-containing molecule. |
From your bench to your data package
| Step | What happens | What you do |
|---|---|---|
| 1. Scope | A 30-minute call fixes the panels, the number of candidates and the material required. You receive a written quotation before anything ships. | Tell us what you are trying to decide. |
| 2. Ship | We confirm the exact amount and concentration per candidate and handle import and cold chain from your door. | Send purified candidates, 20 µg per candidate and panel. |
| 3. Measure | Your candidates run on MACS® Matchmaker against the panels you selected, with within-chip replicates and controls in every run. | Nothing. |
| 4. Deliver | You receive the structured data package one to two weeks after sample arrival, plus a call with the scientist who ran the measurements. | Decide. |
A dataset, not a PDF
Machine-readable One normalized record per candidate Tabular and JSON output with the raw sensorgrams attached, built to feed a model rather than to sit in a slide deck. | Positioned Every candidate against the benchmark Each value reported against the clinical-stage reference distribution, so a result can be acted on without a specialist to interpret it. | Complete Every axis, every candidate Affinity, polyspecificity, matrix behavior, FcRn and Fc gamma receptor profile, reported per candidate and comparable across your whole set. |
Native No labels, no tags, no conjugation Your molecules are measured as they are. Nothing is modified to make the measurement possible. | Frugal Micrograms per candidate Low sample consumption leaves material available for the rest of your program. Conventional plate methods consume tens of micrograms per readout. | Documented The same assays we publish Every panel is described in full on our application pages, with protocols, configurations and controls. |
Three ways teams use the service
AI protein design Measure the axis your models were never trained on You generate more candidates than you can characterize, and public training data covers affinity, not liability behavior in real matrices. We return developability data in a form a model can read. | Antibody discovery Get the liability profile without building the assay You have a short list and no biophysics capacity to spare. Get polyspecificity, matrix behavior and Fc biology on every candidate without buying an instrument or developing a method. | Developability groups Overflow capacity with data you can defend Your queue is full and discovery keeps asking. Use the service for polyreactivity and matrix work, with documented protocols, controls in every run and raw data included. |
What we measure, and what we do not
What the service answers- Whether a candidate binds non-specifically, and how that compares to clinical-stage antibodies - How it behaves in undiluted serum and plasma rather than in buffer - Whether FcRn binding will let pharmacokinetics translate across the species in your studies - What the Fc gamma receptor engagement profile looks like - How every candidate in your set ranks against every other one on each of those axes | What it does not answerWe measure molecular interactions. We do not measure thermal stability, aggregation, solubility, viscosity at formulation concentration, charge heterogeneity or chemical degradation, and we will tell you so rather than stretch a claim. We also do not design proteins and we do not express them. That is deliberate. A measurement layer is worth more when it has no stake in the molecules it measures. The service is for research use only. |
Questions we are usually asked first
Send us three candidates
Start with a pilot. Three of your molecules, the panels you choose, the full data package with every candidate positioned against clinical-stage antibodies, and a call with the scientist who ran them. You will know whether your lead is cleaner than the drugs already on the market, or whether it carries the liability that ended the ones that are not.