Cortex Virtual ADMET

Boost your leads’ success rate!

Cortex's advanced deep learning system excels in predicting the biological, chemical, and physical properties of small molecules. Our comprehensive services span every facet of preclinical drug discovery, from initial hit identification to lead optimization.

By first screening your hits or analogs with our system, you significantly reduce the risk of drug candidates failing due to compound toxicity or inadequate ADME properties, streamlining the lead optimization process. This approach not only reduces the number of back-and-forth testing iterations but also leads to significant time savings and minimizes the need for costly experiments.

Just provide us with a list of SMILES or equivalent data, and we'll deliver a comprehensive report containing all relevant ADMET predictions for these compounds.

Predicted properties

We consistently expand our repertoire of predicted properties and can accommodate specific requests based on the availability of requisite training data.

  • Physicochemical
    • Log P
    • Solubility in water
    • Solubility in DMSO at 10 mM
  • ADE
    • Solubility in FaSSIF
    • PAMPA permeability
    • Caco-2 A→B permeability
    • Fraction unbound in plasma
    • Half-life in plasma
    • Half-life in liver microsome
  • Toxicity
    • Cytotoxicity in various cell lines
    • Genotoxicity
    • hERG inhibitor
  • Metabolism
    • CYP1A2 antagonist
    • CYP2A9 agonist
    • CYP2C9 antagonist
    • CYP2C10 antagonist
    • CYP2D6 antagonist
    • CYP3A4 agonist/antagonist
    • CYP19A1 antagonist
    • AhR activator
    • CAR agonist / inhibitor
    • PXR agonist
    • CSTO1 inhibitor
    • ALDH1A1 inhibitor

Example

Here is an example of our in silico ADMET reports (click to enlarge):

In this example, we exclusively display metabolic and toxicity targets for which the molecule is predicted to be active. Predictions characterized by low confidence levels are deliberately excluded. Alternatively, you have the option to request the complete set of predictions. Naturally, the data is also accessible in tabular form, structured to align with your preferred format.

Confidence assessment

Crucially, the report furnishes a quantitative measure of the neural network's confidence level for each unique prediction and molecule. The neural network is not solely trained to make precise property predictions but also to gauge the likelihood of their accuracy, enabling it to generate dependable probability estimates of its own confidence.

This empowers you to make informed decisions regarding data interpretation. Our dedicated team is readily available to offer assistance and provide additional insights into each prediction, including details about the training data properties or the neural network's measured prediction accuracy.

Pricing

Our pricing model is designed to cater to projects of all sizes, offering an appealing structure for both small and large-scale endeavors. We will gladly provide a customized quote tailored to your specific requirements.

Contact Us