R&D Productivity Driver

Prous Institute Symmetry℠ integrates computational tools and methods aiming to replicate all the processes through which new drugs are discovered, developed and approved.

Symmetry enables the generation of new research hypotheses and elucidates the efficacy and safety profile of small molecules through a wide range of predictive models, including mechanism of action and phenotypic models, toxicity and human adverse effects. It is a practical tool to increase efficiency.

Symmetry has been designed as a server installation offering scientists at multiple research sites a collaborative environment for drug discovery and safety evaluation. API or web services can also be made available to facilitate integration with customers IT infrastructure.

The platform applies advanced machine learning techniques to a variety of structural features and physicochemical properties of small molecules to provide quality predictions about biological effects. Available Symmetry algorithms include binary classification for active/inactive data sets, meta-classifiers to achieve consensus predictions for sets of binary models and multi-label learning that yields a ranking and probabilistic estimates of the possible outcomes.

We draw upon 50 years of experience in creating and managing biomedical knowledge to create quality data sets comprised of small molecules and associated biological properties. Sources include patents and journals, congresses, data from governmental, academic and commercial partnerships, public access databases and experimental studies at Prous Institute. New data sets are continuously generated as information becomes available and are used to validate and update Symmetry models, as well as to create new models.

Key Applications and Benefits

  • Provides a multidimensional view for decision support in the design, optimization and prioritization of compounds prior to synthesis, saving time, costs and minimizing risk.
  • Elucidates new mechanisms of action and therapeutic applications for small molecules and facilitates emerging target assessment.
  • Identify potential off-target interactions and use results in decision support to guide experimental assays.
  • Alerts for preclinical toxicology liabilities and safety end points of critical importance to regulatory authorities.
  • Anticipates the strengths and weaknesses of competitor compounds by assessing their undisclosed in silico pharmacological and toxicological profiles.
  • Enables rapid high throughput virtual screening.
  • Allows custom-built models with proprietary data.
Through Symmetry’s intuitive interface, users can easily load and predict the biological effects of small molecules as shown.

With Symmetry’s advanced model building capabilities users can construct powerful predictive models with their proprietary data.

• Build custom models with in house data alone or in combination with Symmetry data sets.
• Multiple model building options and algorithms permit flexibility in finding optimal settings.
• Automatic internal validation and external validation reports to assess and select best performing models.
• Compare statistics for multiple models side-by-side.
• Group models into suites and obtain consensus outcomes.