Wide range of state-of-the-art features that enable:

• Understanding possibilities and limitations of data
• Building 'best possible' models
• Capture best practices in model building
• Build and validate multiple models interactively
• Extreme ease-of-use

All of this power and expertise organized into guided modeling and prediction Workflows and in a convenient layout that minimizes repetitive steps and maximizes efficiency.

Learn more about Structure Processing : Data Exploration Algorithms : Model Building Algorithms

Metrics to assess confidences at every step of model building and predictions to help in decisions such as:

• Ascribe confidence to property predictions
• Applicability of models to chemical series
• Model expansion or model localization
• Selecting suitable algorithms and parameters
• Designing 'global-local' models through Nested and Consensus modeling

All of this power and expertise organized into guided modeling and prediction Workflows and in a convenient layout that minimizes repetitive steps and maximizes efficiency.

Learn more about Prediction Confidence Metrics : Applicability Metrics : Model QA Metrics

Models built using Sarchitect allow packing various types of information with them. This enables:

• Communication of vital model information to users
• Archiving modeling decisions and findings
• Model expansion and 'localization'
• Improvement of model accuracies over time
• Computation of confidence metrics

No more blackboxes!

Learn more about Model information for Packaging

Several ADME and Toxicity models built on carefully curated data are available in Sarchitect. Our Models are open, they come packaged with the underlying training set, structures and data. We also include important statistical information as part of the model metadata. Both are used in the critical assessment of prediction space.

Use our models or build your own, with our data!

Learn more about Strand ADME Models : Strand Toxicity Models

A new paradigm of in silico lead optimization – create or edit a structure and make predictions, real-time. This feature enables studying the variation of properties with:

• Structure modifications
• R-group enumerations
• Diverse compound libraries

No more waiting!

Learn more about Compound Profiling