Propersea (Property Prediction)

Propersea predicts a wide range of molecular and physicochemical properties for small molecules, such as melting and boiling points, density, solubility, polarizability, and more. Propersea can also predict IUPAC names using a machine learning model developed by the PSDS team. It employs various algorithms, including RDKit, semi-empirical quantum methods, Bayesian regression trees, and transformer neural networks. Results include predicted values, confidence intervals (for Bayesian models), and comparisons to molecules in the training set. Propersea excels with organic compounds but performs less reliably with inorganics and organometallics, as indicated by lower reliability metrics. These property predictions play a vital role in physical sciences, helping researchers save time and resources while accelerating the development of innovative materials, reactions, treatments, and others. Disclaimer - Propersea contains calculated rather than experimental properties.