The Meaning and Mastery of Metadynamics
International computational scientist Michele Parrinello speaks at PNNL
In his technical talk, titled "Proteins in Motion," Italian computational scientist Michele Parrinello discussed how studying a protein's structure can expand our understanding of how proteins operate and work. He focused on the computational chemistry technique of metadynamics, which allows efficient sampling and permits an accurate reconstruction of the free energy landscape. In metadynamics, sampling is accelerated by a time-dependent bias that acts on a small number of collective variables. He illustrated that an appropriate choice of these collective variables leads to some spectacular success.
Also, Parrinello described the novel variant reconnaissance metadynamics, in which machine learning techniques are used to find the collective variables automatically and reconstruct the free energy landscape.
Parrinello is a professor of computational sciences at ETH in Zurich, Switzerland, and a professor of computational chemistry at Scoula Normale in Pisa, Italy. He has made many contributions to the fields of materials science, catalysis, complex chemical reactions and hydrogen-bonded systems.
The talk is part of the Laboratory Director's Distinguished Lecture Series at PNNL. The series features high-profile science, technology and policy leaders who share their insights on the current and future state of science and engineering.
The lecture was held on Tuesday, October 18, at the Battelle Auditorium. A well-attended reception for Parrinello followed, providing the speaker a chance to converse with members of the national lab and the public.