Welcome to CASCADE-ChemicAl Shift CAlculator using DEep Learning

on Jul 3, 2019

CASCADE is a stereochemically-aware online calculator for NMR chemical shifts using a graph network approach developed at Colorado State University. This provides chemical shift predictions trained to achieve DFT-levels of accuracy within seconds: the average error against 5,100 testing experimental shifts is 1.43 ppm for ¹³C. The CASCADE methodology, training data used, and typical performance statistics are all described in the following publication: Chemical Science. The underlying code and training data are also made available through GitHub

The current online release is limited to the prediction of ¹³C chemical shifts with experimentally corrected results. Similar ¹H results are predicted using the DFT based neural network. Query structures can be input via graphical interface or SMILES input. This launches 3D-structure generation, conformational search, and a prediction of chemical shifts for the Boltzmann-weighted ensemble. More features are under development. Use the Predict tab to explore more.

This website is currently being tested internally. Stay tuned for more updates!

CASCADE is open to the world!

on Jul 19, 2019

CASCADE is now open and being tested worldwide. Please feel free to email us any feedback or comment.