Simple to use, accurate, and
reliable, SciLogP
Ultra is a major
advance in predicting LogP, the
octanol/water partition
coefficient, for "neutrally
charged" organic compounds.
SciLogP Ultra Features
- Based on 2D E-State descriptors and neural net approach, SciLogP
Ultra gives R² = 0.95 and mean error of 0.3Log units on
3,800 new and diverse compounds outside the training and test sets. This
is unparalleled for any LogP predictor.
8,905 compounds were used from the BioByte Starlist to construct SciLogP Ultra in MW range from 30 to 900.
Non-Fragment Approach allows you to process thousands of molecules in just minutes, without loss of LogP values due to missed fragments.
Built-in error analysis indicates reliability of predicted LogP.
Built-in similarity search capability using 8,000 compounds from
BioByte's Starlist.