SEMINAR: Full Stack Molecular Modeling: Building Open Infrastructure for Interoperable, Reproducible, and Extensible Physical Simulation

SEMINAR: Full Stack Molecular Modeling: Building Open Infrastructure for Interoperable, Reproducible, and Extensible Physical Simulation

Seminar Abstract 

Molecular simulation addresses scientific challenges in a wide variety of fields, ranging from drug design and protein biology to catalysis and solar energy capture. However, the power of our available tools, whether for molecular mechanics, chemoinformatics, or quantum chemistry, is too often limited by practicalities, such as the difficulty of compiling and deploying complex software, incompatible input and output formats, poor interoperability, and the idiosyncratic complexities of distributed computing. To reduce the barriers to computational modeling, the Autodesk Life Sciences group is building open source, cloud-based software that makes it easier to create, deploy, and use molecular modeling protocols. In this talk, I'll describe the three main components of our group's software stack: the Molecular Design Toolkit, an abstract Python API for molecular modeling methods; the Molecular Workflow Repository, an open collection of reproducible and extensible modeling protocols that can run anywhere; and the Molecular Simulation Toolkit, a collection of easy-to-use, web-based graphical applications that give a wide range of scientists access to complex, validated modeling protocols. In building this combination of domain-specific modeling tools, validated workflows from domain experts, and modern software infrastructure, we're working to put nanoscale modeling on a path to wider usage, easier development, and a faster pipeline from research to industry.

Seminar Speaker(s) 

Aaron Virshup
Principal Research Scientist , Autodesk

Aaron Virshup is a Principal Research Scientist in the Life Sciences group at Autodesk Research in San Francisco. He works to enable and democratize nanoscale design by combining academic research with modern computer-aided-design infrastructure. Before coming to Autodesk in 2015, Aaron was a postdoc at Duke University, where, with Professors Weitao Yang and David Beratan, he created and applied algorithms for chemical library design to problems ranging from drug design to photochemistry, and developed a novel molecular pathway theory for excitonic Dexter-type energy transfer. Before that, Aaron received his PhD in Physics from the University of Illinois under Professor Todd Martinez, where he investigated and developed methods to model light-activated quantum dynamics in complex chemical environments.

Date 

May 31, 2017 -
12:00pm to 1:00pm

Location 

90-3122

Host