Function measured at library scale

Most discovery methods tell you which antibodies survived a selection. Deep Screening tells you how tightly every one of them binds — quantitative binding paired to full-length sequence, across millions of variants and across many conditions, in a single experiment.

150M ANTIBODIES / EXPERIMENT
6M UNIQUE, 12× REPLICATED
K_D QUANTITATIVE READOUT

Measurement, not just selection

Selection-based methods recover the variants that survive. Deep Screening adds the number they leave out — how tightly each one actually binds.

Phage display and other display methods are powerful ways to find antibodies. Deep Screening sits alongside them and reads out something they can’t: a quantitative binding value for every variant in the library.

With Deep Screening, hundreds of millions of antibodies are assayed in parallel, and each one is paired to its full-length sequence. The result is a dense, unbiased dataset, measured on the actual target.

Because the readout is a measurement rather than a survival signal, the same library can be screened across different conditions — pH, competing ligands, species orthologues — and compared variant by variant. That is what turns conditional behaviours from lucky findings into something you can engineer.

To learn more, read the paper: Deep Screening published in Nature Biomedical Engineering

Multiple properties across parallelised experiments

When every variant is measured rather than just selected, several drug-relevant properties can be read out from the same screen, instead of chased one assay at a time.

Target specificity: Resolve close homologues — bind one family member while sparing related ones — across the full library.
Species cross-reactivity: Human and tox-species binding measured in the same screen. Translational confidence built in early.
Polyreactivity: Off-target binding mapped at the screen — developability assessed alongside potency, not after it.
Affinity: Push from high pM toward single-digit pM / fM, with the full affinity landscape in view.

Conditional binding

Conditional binding is the hard problem where measuring across conditions matters most. Done well, it widens the therapeutic window, maximising effect at the target while sparing healthy tissue. Our focus is on engineering conditional binders, leveraging Deep Screening to control when, where, and how an antibody engages its target.

SENSE THE ENVIRONMENT

Binders that detect the distinct biochemistry of a target tissue — pH, proteases, cofactors — and respond only there. Tuning the sensor needs binding measured across those conditions, not just in one.

SWITCH BINDING ON AND OFF

A masking domain or conformational lock that keeps the molecule inert in circulation and releases on trigger. The on- and off-states have to be measured and balanced, not assumed.

TARGET SELECTIVELY

Full engagement unlocked only when the trigger is present, optimised against every state at once so systemic safety and local efficacy are designed together rather than traded off.

Built for complex protein engineering

We partner with pharma and biotech on engineering challenges that resist conventional methods -- conditional binding, fine selectivity, affinity at the limit -- where measuring across conditions makes complexity tractable

Partner with us

Technology