AI-based platform for small molecules liability profiling

Drug off-target interactions are one of the main reasons of candidate failure in the drug discovery process. Anticipating potential drug’s adverse effects in the early stages is a mandatory step to minimize risks of candidate failure.

ProfhEX is an AI-driven suite of 46 OECD-compliant machine learning models able to profile small molecules on 7 liability groups relevant for drug discovery campaigns, including: cardiotoxicity, neurotoxicity, gastrointestinal toxicity, endocrine disruption, renal, pulmonary, and immune system toxicities.

To estimate a liability profile, upload molecule(s) in SMILES format: ProfhEX will provide downloadable target-inhibition activities together with a summary report. Available models can be browsed in the Models section.


Deep Learning


Models Selection



Submit Your File

To run ProfhEX, either paste a SMILES string ( for an example)
upload a .csv,.txt,.smi file containing one SMILES per row ( for a template)

Once calculations are completed, results can be analyzed and downloaded.

Please note that the maximum number of processed molecules caps at 100 (*) (exceeding rows are ignored).

Powered by


Available models can be selected in the drop-down list or browsed by liability group in the treemap visualization. The report displays key model quality information:

  • Training set size and pK property distribution
  • External validation performance
  • Internal validation evaluation

* To use this tool on Edge and Firefox you must allow cross-site tracking cookies or add the site name to the tracking prevention exceptions in the privacy settings; on Chrome you do not need to.



We will reply as soon as possible

Web Site

Always available for you