Solute descriptors have been widely used for modeling various chemical transfer processes through linear free energy relationships (LFER); however, the lack of measures to obtain these descriptors accurately and quickly for new organic chemicals has greatly limited their applicability. In this research, models (PaDEL-DNN) that require only SMILES of chemicals to satisfactorily estimate LFER descriptors were built using deep neural networks (DNN) and the PaDEL molecular representation. The newly estimated LFER descriptors demonstrated good performance in modeling storage-lipid/water partitioning (log<i>K</i><sub>storage-lipid/water</sub>), bio-concentration factor (BCF), aqueous solubility (ESOL), and hydration free energy (freesolve). Then, assuming that the accuracy in the obtained values of widely available properties, e.g., logP (octanol-water partition coefficient), can calibrate estimates for less available but related properties, we proposed logP as a surrogate metric for evaluating the overall accuracy of the estimated LFER descriptors. When modeling the above four chemical transfer processes with selected chemicals whose estimated LFER descriptors were deemed “accurate” by the surrogate metric, we observed around 0.1 log unit lower errors. The interpretation of the PaDEL-DNN models revealed that, for a given test chemical, having several (around 5) ‘similar’ chemicals in the training dataset is crucial for accurate estimation while the remaining less similar training chemicals provide reasonable baseline estimates. Lastly, LFER descriptors for over 2800 persistent, bio-accumulative, and toxic chemicals were reasonably estimated by combining PaDEL-DNN with the surrogate metric. Overall, the PaDEL-DNN/surrogate metric approach and the newly estimated descriptors will greatly benefit chemical transfer modeling.

Predicting solute descriptors for organic chemicals by a deep neural network (DNN) using basic chemical structures and a surrogate metric

ACS Spring 2022

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.