Predict enzyme interactions
with DeepMolecules

DeepMolecules offers state-of-the-art machine learning models for predicting substrates for enzymes and transport proteins, as well as enzyme kinetic parameters (kcat, KM).

Free and open to all users — no login required.

How it works

Choose your prediction task

Select a model below to predict enzyme-substrate interactions or kinetic parameters from sequence data.

ESP

Enzyme-Substrate Pair Prediction

Predict whether a given enzyme will accept a specific substrate molecule, powered by our ProSmith model with improved accuracy.

Start prediction
TurNuP

kcat Prediction

Predict the catalytic turnover number kcat (s-1) for enzyme-substrate pairs from amino acid sequences and SMILES.

Start prediction
KM

KM Prediction

Predict the Michaelis constant KM (mM) — the substrate concentration at which the reaction rate is half-maximal.

Start prediction
SPOT

Transporter-Substrate Pair Prediction

Predict whether a membrane transporter protein will transport a given metabolite substrate across the cell membrane.

Start prediction
Coming soon
FCKcat

Mutation Effect Prediction

Predict the fold change in kcat caused by a mutation, by comparing the wild-type and mutant enzyme sequences.