By clicking “Accept”, you agree to the storing of cookies on your device to enhance site navigation and analyze site usage.

Preclinical partnerships

From target to investigational new drug (IND) application.
FTE-based, upfront, monthly fees, milestone payments and royalties.
Experimental validation via our CROs.
One to two years.
Acellera ACEMD preview code

Our approach

We focus on preclinical drug discovery where we aim to accelerate discovery and reduce experimental costs using molecular simulations and machine learning.
An in-house GPU cluster so that your data will not leave our facilities.
Carbon neutral since 2018 for our electricity sources.
Strong know-how in computer simulations and machine learning. See our publications in top-tier journals.
Proven collaborations with pharmaceutical companies.
We work with established CROs to deliver experimentally validated molecules.
9.4/10 rating in Science Exchange.
Cryptic Pockets and Fragments


Identify the different pockets of your target, assess their drugabbility and reveal cryptic sites with mixed solvent molecular dynamics (CrypticScout). 

Gather all the available knowledge for your target (structures, affinity and selectivity data) using AceProfiler.

Identify the binding mode of the known actives using short molecular dynamics simulations (FragmentScout.).

Obtain mechanistic insights for your target (flexibility, different pocket conformations, etc.) using our in silico conformational assay (long molecular dynamics simulations)


Screen millions of purchasable compounds against your target and pocket of choice using our award-winning docking platform. Obtain accurate affinity predictions for your compounds with our dynamic undocking (DUck) protocol and our deep learning predictor (KDeep). 

We can also run the assays for the best compounds thanks to our CROs.
Dashboard mockup
small molecule table


Enhance your hit’s affinity leveraging our machine learning tools (KDeep and AceGen) and our molecular dynamics protocols (Dynamic Undocking and Relative Binding Affinity prediction).


Improve the profile of your lead compound in different aspects such as solubility, toxicity, selectivity or activity trough our generative platform and our re-trainable, deep learning predictor.

Leveraging public and in-house data, we will train a deep learning model to predict different properties of your interest and select the compounds with the best predicted profile.

This approach granted us the second place in solubility prediction in the logP challenge.
kdeep input

Delivering High-Quality Molecules Through Our Platform

Through CRO partnerships we have access to biochemical and cell-based assays across target classes (GPCR >2750 assays, Kinase/Enzyme >1000 assays, Ion Channel >400 assays).

By conducting experiments, we can compare the results of computational predictions with real-world observations. This helps us to determine the accuracy and reliability of the predictions and provides feedback to continuously improve our tools. At the same time, experimental validation filters and selects our molecules to deliver high-quality drug candidates in our drug discovery campaigns.

Want to know more?

Find out how we can help.
Contact us