Our softwares


TOP SERVICES

QSAR Studies, new tools and more.

24/7 SUPPORT

Please contact us for support.

FREE QUOTE

Contact us to request trials or free quote.

NEWS

Take a look in our activities, events and news.

Artificial Intelligence and Advanced Machine Learning Algorithms

Artificial intelligence and advanced machine learning algorithms are used for toxicity assessment with structure-toxicity relationship (STR) probability mapping.

  • Modeling by statistical and artificial intelligence algorithms: random forest model (RF), k-nearest neighbors algorithm (k-NN), and deep learning models using datasets containing rigorously curated structures
  • Structure-toxicity relationship (STR) probability mapping indicates the fragments more related to the absence/decrease of toxicity (green) or presence/increase (red), useful for hypotheses and mechanistic interpretations
  • The interactive visualization tools help the user know when to apply expert knowledge that could potentially refute a (Q)SAR prediction and how to accomplish this

Visual AD Inspection®

The applicability domain (AD) is defined by the chemical structure space and the toxicological response encoded by the developed model to make new predictions with a given reliability (a defined domain of applicability, OECD Principle 3). Our visual AD inspection is used to establish the scope and limitations of the models. Basically, new chemicals must be reasonably similar to training set compounds or a prediction cannot be accepted.

Results Based on the Most Predictive Models Aligned with the (Q)SAR Validation for Regulatory Purposes

  • A standardized report is generated to ensure that the results are consistently documented, transparent, and complete
  • Mechanistic interpretation
  • Easy visual inspection of the accuracy, confidence, and applicability domain
  • The user has direct support from our scientists and developers

OECD Principles of (Q)SAR Validation for regulatory purposes