Special Colloquium
Yang Chen
Harvard University
Delineating Protein Transportation Processes: Hierarchical Models for Single-Molecule Data
Abstract: Single-molecule experiments investigate the kinetics of individual molecules and thus can substantially
enhance our understandings of various organisms. Analyzing data from single-molecule
experiments poses a number of challenges: (a) the inherent stochasticity of molecules is usually
buried in random experimental noise; (b) single-molecule behavior can be highly volatile. For
both of these reasons, replicated experiments are usually required. In order to combine information
from replicated experiments while accounting for the heterogeneities among experimental
replicates, we introduce a Bayesian hierarchical model on top of an experiment-level hidden
Markov model, where the hidden Markov model has been widely adopted to analyze each experimental
replicate individually in previous studies. We apply the proposed model to three data
sets obtained from experiments aimed at unveiling mechanisms underlying protein transportation
– a biological process vital for the proper functioning of cells. Our statistical results enable us
to give a comprehensive picture of the protein transportation mechanism and provide a general
framework for rigorous statistical analysis of experimental replicates from single-molecule experiments.
Please refer to the published manuscript of this work for details, which is available at
http://scholar.harvard.edu/yangchen/publications.
Thursday January 12, 2017 at 3:00 PM in SEO 636