Robust QSAR models enable computational and medicinal chemists to accurately predict the biological activity and ADMET properties of new compounds. This saves time and resources in drug design projects, by enabling researchers to prioritize the best molecules to make.
Field QSAR: activity prediction and model interpretation
Field QSAR uses Cresset 3D descriptors, based on Cresset field technology, to provide a global view of your SAR data, and gives you both prediction and interpretation. It is very efficient at modeling specific ligand-protein interactions to predict biological activity at the target of interest. Visual feedback helps you identify the common favorable and unfavorable features within your molecules so that you can further improve your designs.
Machine Learning models: predict activity and ADMET properties
Flare offers a choice of robust and well validated machine learning methods to build regression and classification models of biological activity and ADMET properties.
Quantitative regression models are suitable when the biological activity data are real values such as pKi or pIC50. Classification models work well for qualitative biological data or data expressed as activity ranges (e.g., % inhibition).
In Flare, you can choose to run individual models or run them all and let Flare pick the best. Activity or category for new compounds can also be predicted using consensus QSAR, to prioritize the molecules all models agree would be a good idea to make.
Models can be calculated using a variety of descriptors, including Flare’s 3D descriptors, modeling the shape and electrostatic character of aligned molecules, fingerprints and descriptors from the RDKit, and custom 3D/2D descriptors.
