Quantifying Mechanical Stress Distribution within Slide-ring Gel
Zi Wang, graduate student
Steve Craig, advisor
Wednesday, April 25, 2018 - 2:00pm to 3:00pm
Location: French Family Science Center 3232
Avery, Meg

Solvation free energy is an important thermodynamic quantity in biological reaction and drug design, and to calculate it accurately and efficiently is highly desirable. In this work, by constructing a database from low-level combined quantum mechanical and molecular mechanical (QM/MM) MD simulations, a machine learning model from gradient boosting (GB) algorithm was built to achieve high-level QM/MM accuracy at low-level QM/MM computational cost. Due to the insufficiency of the database, we applied two checking models that are not kernel based to efficiently distinguish high-dimensional data and extract new configurations. Because of high efficiency of gradient boosting, the solvation free energy was then obtained from MD simulations with QM/MM-GB updated on-the-fly and checking model updated as well.