UW Data Science Seminar: Nathan Baker

Wednesday, April 26, 2017 3:30 Johnson 102

Nathan Baker, Director of the Advanced Computing, Mathematics, and Data Division at Pacific Northwest National Laboratory (PNNL) and a Visiting Faculty member at Brown University, will be presenting "Uncertainty in Biomolecular Solvation" at this week's Data Science Seminar. The Data Science Seminar is free and open to the public.

Abstract

Solvation-related interactions strongly influence a wide range of biomolecular processes. However, both our models and our information for parameterizing those models are imperfect. This talk will describe strategies for quantifying the uncertainty in biomolecular solvation models and optimizing those models to provide the best possible accuracy and performance. The first part of the talk will describe generalized polynomial chaos methods for quantifying solvation energy uncertainty due to conformational noise and errors in atomic charge and radius parameters. The second part of the talk will outline Bayesian methods for addressing model uncertainty through statistical aggregation of predictions from multiple models.

Guest Seminar: Andrew Hufton

Tuesday, Apr. 18, 1:30 p.m., Smith Hall 105

The UW eScience Institute's Repoducibility and Open Science Group is hosting Andrew Hufton, managing editor of scientific data Nature Research. Hufton's talk is entitled "Beyond supplementary material: Sharing data effectively through repositories and data journals."

ABSTRACT
The Nature Research journals understand that effective data sharing supports reproducibility and can increase the impact of published works. Indeed, our policies have long recognized that data sharing is a fundamental part of research publication. The increasing complexity and size of research datasets, however, poses challenges for scientists who wish to share their data in a reusable and transparent manner. Based on my experience at Scientific Data, an open-access data-focused journal from Nature Research, I will provide tips on how researchers can share their data in an effective manner that promotes reuse, supports the credibility of their research, and ensures they get proper credit. This will include advice on writing better data-rich papers, the basics of presenting datasets in a useful manner, and tips on how to find the right repository for your data. I will also explain Scientific Data's editorial policies and share some of our experiences peer-reviewing and publishing data so far.

BIO
Andrew is responsible for the editorial policies of Scientific Data, in consultation with the Honorary Editor and Advisory Panel, and works with the Editorial Board to ensure a fair and thorough peer-review process for all submissions. Andrew received his PhD from Stanford University in 2006, and did postdoctoral work at the Max Planck Institute for Molecular Genetics in Berlin. His research included topics in developmental genetics, computational biology and genome evolution. Before joining Scientific Data, Andrew worked as an Editor at Molecular Systems Biology.