Candidates with evidence
Sequence and structure comparison, functional-site checks and Foldiff and ThermoMPNN predictions with explicit disagreements.
STRUCTURE → HYPOTHESIS → EXPERIMENT
Fewer guesses.
Better evidence
for your experiment.
Choose your first enzyme stabilization variants. Compare structures, predictions and reported measurements, then test your hypotheses in the laboratory.
Try a worked example ↗AlphaFold DB · UniProt · FireProtDB
Computational evidence alongside experimental results.
01 / ENZYME STABILITYValidate predictions in the laboratory
WHAT YOUR LAB GETS
From structure to measurement plan
Your first variant series
Select single substitutions, a WT control and replicates. Review combination and primer designs with stated limits and exports.
Measurements and next steps
Enter results or import CSV, compare with WT and check activity separately. For pH, view a curve and observed maximum.
WORKED EXAMPLE
Subtilisin and thermitase
The saved example starts with 150 differences and applies structure and functional filters. Eight candidates remain for experimental selection; the predictors disagree on some variants.
Explore the example ↗Your protein↓Comparison and checks↓Candidates with evidence↓Laboratory series↓Measurements → next step
SCOPE
A research tool for enzymes
When it helps
When you know the protein, have a suitable structure and need a reasoned choice of variants for a first experiment. The workspace also compares homologs and plans activity measurements across pH values.
What the laboratory establishes
A substitution’s actual effect, retained activity and transfer to your conditions. Predicted ΔΔG cannot be converted into a guaranteed temperature or activity gain.
FOR LABORATORIES
One enzyme.
A clear success criterion.
For a pilot, prepare the protein, goal, application conditions, construct and assay method. Start with a public entry and worked example.
Current research beta
Calculations, series planning and measurement analysis are available in the workspace. Experiment series stay in your browser. Download CSV and JSON; manually restore JSON on another device.
Open workspace ↗Discuss a joint pilot via @oligoplan_bot. Scope, timelines and fees are agreed separately. Confidential data require a private laboratory account.
Frequently asked questions
Does the service make an improved enzyme?
It proposes candidates and evidence for testing. Constructs are made and tested in the laboratory; a prediction alone cannot confirm an effect.
Do I need to run AlphaFold?
Supported entries use existing AlphaFold DB structures. Predicting structures for other sequences is on the roadmap.
Can I bring experiment results back?
Yes, manually or through CSV. The service calculates means, spread and differences from WT. Uploaded measurements do not automatically retrain the model.