Skip to content
Free feasibility assessment

Case studies

Do computationally designed antibodies actually bind?

IL-7Rα

  • 15candidates designed by computation, produced and tested
  • 7confirmed antibodies, measured by SPR
  • 47 %hit rate, against a few per thousand in classical screening
  • 1.4 nMaffinity of the best clone in the panel

Our own designs, on IL-7Rα, measured by SPR in an independent laboratory.

PD-L1

  • 15candidates designed by computation, produced and tested
  • 6confirmed antibodies, measured by SPR
  • 40 %hit rate, against a few per thousand in classical screening
  • 13 nMaffinity of the best clone in the panel

Our own designs, on PD-L1. Stark et al., 2025.

PMVK

  • 15candidates designed by computation, produced and tested
  • 8confirmed antibodies, measured by SPR
  • 53 %hit rate, against a few per thousand in classical screening
  • 6.1 nMaffinity of the best clone in the panel

Our own designs, on PMVK. Stark et al., 2025.

The method and the services

  • Our technology

    We start from a validated antibody, already screened for liabilities, and design the CDR loops that grip your target

  • All services

    Nanobody, antibody, miniprotein binder and peptide discovery, de novo design and humanization

Start designing better antibodies

We design developable antibodies using AI in under 6 weeks, wet-lab validated in 9 weeks, from your antigen sequence. No immunisation, no phage display, no animal.

Receive a feasibility study on your antigen - for freeTalk to us