These are preliminary reports that have not been peer-reviewed. They should not be regarded as conclusive, guide clinical practice/health-related behavior, or be reported in news media as established information. For more information, please see our FAQs.
SAMPL6 Challenge Results from pKa Predictions Based on a General Gaussian Process Model
preprintrevised on 26.09.2018, 00:42 and posted on 26.09.2018, 12:58 by Caitlin C. Bannan, David Mobley, A. Geoff Skillman
A variety of fields would benefit from accurate pKa predictions, especially drug design due to the affect a change in ionization state can have on a molecules physiochemical properties.
Participants in the recent SAMPL6 blind challenge were asked to submit predictions for microscopic and macroscopic pKas of 24 drug like small molecules.
We recently built a general model for predicting pKas using a Gaussian process regression trained using physical and chemical features of each ionizable group.
Our pipeline takes a molecular graph and uses the OpenEye Toolkits to calculate features describing the removal of a proton.
These features are fed into a Scikit-learn Gaussian process to predict microscopic pKas which are then used to analytically determine macroscopic pKas.
Our Gaussian process is trained on a set of 2,700 macroscopic pKas from monoprotic and select diprotic molecules.
Here, we share our results for microscopic and macroscopic predictions in the SAMPL6 challenge.
Overall, we ranked in the middle of the pack compared to other participants, but our fairly good agreement with experiment is still promising considering the challenge molecules are chemically diverse and often polyprotic while our training set is predominately monoprotic.
Of particular importance to us when building this model was to include an uncertainty estimate based on the chemistry of the molecule that would reflect the likely accuracy of our prediction.
Our model reports large uncertainties for the molecules that appear to have chemistry outside our domain of applicability, along with good agreement in quantile-quantile plots, indicating it can predict its own accuracy.
The challenge highlighted a variety of means to improve our model, including adding more polyprotic molecules to our training set and more carefully considering what functional groups we do or do not identify as ionizable.