Platform
Our lead optimisation platform unlocks high-performing molecules, often with non-intuitive designs.
Building Blocks
The EVA™ Platform
Lead Candidates
Diverse binders
with regards to epitope coverage and affinity
Linkers
with varying lengths
and flexibilities
Half-life extension domain
T-cell
Engagers
ADCs
NK-cell
Engagers
TCR-based
Engagers
Agonists
Cell Therapies
Radioconjugates
Multispecifics
We can work with any modality, providing there is a high-throughput, disease-relevant, cell-based assay.
What makes this possible?
Closed-loop discovery
Our discovery engine tightly integrates machine learning (ML) with automated functional screening in a closed loop.
Proprietary ML-grade data
Data from both functional and developability assays fuels a powerful search technique called Multi-Objective Bayesian Optimisation. This form of ML enables us to rapidly evaluate large numbers of multispecifics (>10^5 designs) for both function and developability.
Speed
The ‘design-build-test-learn’ cycle takes <6 weeks. Each antibody panel consists of 2,300 multispecifics. Mass modular cloning combined with in-house sequence verification allows for rapid construct assembly. Each design is expressed in a mammalian host, purified and tested across a suite of in-depth assays for both function and developability.
Flexibility
Our platform is both modality and format-agnostic.
Building Blocks
The EVA™ Platform
Lead Candidates
Dry Lab
In silico design
Antibodies are designed computationally using insights extracted by ML from previous cycles.
Multi-objective optimisation
Proprietary ML-grade experimental data build and refine predictive models, significantly increasing the number of antibodies evaluated.
<6 weeks
State-of-the-art automation lab
Antibody production
Up to 2,300 multispecifics are cloned, sequence-verified, expressed in mammalian systems, and purified.
Characterisation
Final-format antibodies are evaluated using high-throughput developability and functional (disease-relevant, cell-based) assays.
