Antibody & Biologics · AstraZeneca / Prihoda et al.
Sapiens
Germline-guided antibody humanisation tool proposing mutations that increase humanness while preserving affinity.
What Sapiens is
Sapiens is a computational antibody engineering tool developed by AstraZeneca and Prihoda et al. to support antibody humanisation. It operates on antibody protein sequences to identify modifications suitable for therapeutic development.
During lead optimisation, engineering non-human or candidate antibodies to resemble human antibody repertoires helps minimise immunogenicity in patients. Sapiens addresses this by proposing the minimal set of germline-guided mutations required to raise sequence humanness without collapsing antigen binding affinity, having undergone validation on clinical-stage molecules.
How it works
Sapiens uses paired heavy and light chain amino acid sequences to suggest germline-guided sequence alterations. Specific internal neural network architectures, pre-training objectives, and training datasets are not detailed in the available documentation.
Key facts
- Version
- 1.0
- Domain
- Antibody
- Modalities
- Protein sequence
- GPU memory
- 8 GB
What you can use it for
- Antibody humanisation
- Proposing minimal sets of germline-guided mutations to increase the humanness of antibody candidates.
- Affinity-preserving engineering
- Selecting humanising sequence changes designed to avoid the loss of binding affinity.
- Clinical candidate progression
- Evaluating mutations and humanness deltas on candidate molecules heading toward clinical evaluation.
Run it here
No installation and no GPU of your own: fill in the inputs and it runs on our servers. Results stay in your browser unless you sign in.
Runs on On-demand GPU
Runs on our servers with on-demand compute. A first run needs time to load the model; active capacity can be reused and scales down when idle.Run this model
Sapiens
Strengths and limits
Good at
- Identifies a minimal set of germline-guided mutations to improve humanness.
- Designed to prevent the loss of target binding affinity during sequence modification.
- Validated directly on clinical-stage antibody molecules.
Watch out for
- Requires both heavy and light chain sequence inputs.
- Requires an environment with at least 8 GB of GPU memory.
Inputs and outputs
You provide
Heavy chain
Single-letter amino acid code. Whitespace and FASTA headers are stripped automatically.
Light chainoptional
Single-letter amino acid code. Whitespace and FASTA headers are stripped automatically.
You get back
Proposed mutations with humanness delta.
Licence
Commercial use permitted · BSD-3-Clause
The BSD-3-Clause licence allows use in commercial research and products.
Questions
Can I use Sapiens commercially?
Commercial use permitted. The BSD-3-Clause licence allows use in commercial research and products.
How do I run Sapiens without installing anything?
Use the form under “Run it here” on this page: fill in the inputs and press Run model. It runs on our servers on 1× A100 16GB, typically in 4.0s, and a run costs about 28 credits. Visitors without an account get 1,000 credits a month.
What inputs does Sapiens require?
The model requires two protein sequence inputs: the antibody heavy chain and the antibody light chain.
What output does the model provide?
It produces a tabular output of variants listing the proposed mutations together with the calculated change in humanness.
How much GPU memory is required?
Sapiens requires 8 GB of GPU memory to run.
On what type of data has the tool been validated?
The tool has been validated on clinical-stage molecules rather than purely computational benchmarks.
Written from Sapiens's own documentation. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.