← All tools
Property prediction · Cheminformatics

DLKcat

Predict an enzyme turnover number from its sequence and a substrate structure.

Runs the deep-learning half of the DLKcat toolbox, which pairs a graph neural network over the substrate with a convolutional network over the enzyme sequence to predict kcat. It is trained on wild-type and mutant enzymes across many organisms, and is meant for parameterising models rather than for ranking closely related variants.

Documentation ↗ Source code ↗
Input

Configure a small job

POST
The substrate as a SMILES string; DLKcat builds its molecular graph from this.
Carried through to the output for readability; it does not affect the prediction.
More substrate/enzyme pairs to score in the same run, one per line, as name;SMILES;sequence.
Ready
Response

JSON result

{
  "status": "Ready for a request"
}