Date: April, 14 2006

Time: 3:00 p.m.

Location: Centergy One 5186

Speaker(s): Dr. Nicoleta Serban
ISyE, GT

Title: High-Dimensional NMR Spectra Identification for Protein Structure Determination

Abstract:
The motivating application of this talk is protein structure determination using nuclear magnetic resonance (NMR). We analyze the frequency domain NMR data or normal spectra obtained by Fourier transforming n-dimensional time domain NMR signals, which are sums of decaying sinusoids. Most of the current research relevant to our application is developed for two-dimensional NMR frequency data. But NMR spectroscopy data for protein structure determination are currently produced even in $5$ dimensions. Our statistical method takes a novel overall perspective: (1) It incorporates a preliminary step for separating the signal from the background using a method that adapts for sharp changes in the data and non-homogeneous signal; (2) The location, width and the amplitude of the NMR spectra identified above a noise level-dependent threshold are estimated using an iterative algorithm; (3) It detects mixtures of spectra at some confidence level.

Bio:
Before joining Georgia Tech - IsyE in 2005, I received a B.S. in Mathematics and an M.S. in Theoretical Statistics and Stochastic Processes from the University of Bucharest. My doctoral degree is in Statistics earned at Carnegie Mellon University. My primary methodological research stream is related to analysis of multiple processes (e.g. multiple time series). A large portion of my research includes designing tests of hypothesis, nonparametric functional estimation, dynamics and change point analysis of temporal processes. My current applied research is two-fold: computational statistics with direct application to molecular biology and finance informatics. Two biological applications I have investigated are multiple gene expression profiles and three-dimensional structure of proteins using nuclear magnetic resonance. Some of my work in finance/business informatics is relevant to learning the business workflow from audit trial data provided by complex IT systems, identifying micro- and macro-economic factors that determine changes and dynamics in relevant business related variables, and discovering patterns in high-frequency financial data.

Date:	April, 14 2006
Time:	3:00 p.m.
Location:	Centergy One 5186
Speaker(s):	Dr. Nicoleta Serban ISyE, GT
Title:	High-Dimensional NMR Spectra Identification for Protein Structure Determination
Abstract:	The motivating application of this talk is protein structure determination using nuclear magnetic resonance (NMR). We analyze the frequency domain NMR data or normal spectra obtained by Fourier transforming n-dimensional time domain NMR signals, which are sums of decaying sinusoids. Most of the current research relevant to our application is developed for two-dimensional NMR frequency data. But NMR spectroscopy data for protein structure determination are currently produced even in $5$ dimensions. Our statistical method takes a novel overall perspective: (1) It incorporates a preliminary step for separating the signal from the background using a method that adapts for sharp changes in the data and non-homogeneous signal; (2) The location, width and the amplitude of the NMR spectra identified above a noise level-dependent threshold are estimated using an iterative algorithm; (3) It detects mixtures of spectra at some confidence level.
Bio:	Before joining Georgia Tech - IsyE in 2005, I received a B.S. in Mathematics and an M.S. in Theoretical Statistics and Stochastic Processes from the University of Bucharest. My doctoral degree is in Statistics earned at Carnegie Mellon University. My primary methodological research stream is related to analysis of multiple processes (e.g. multiple time series). A large portion of my research includes designing tests of hypothesis, nonparametric functional estimation, dynamics and change point analysis of temporal processes. My current applied research is two-fold: computational statistics with direct application to molecular biology and finance informatics. Two biological applications I have investigated are multiple gene expression profiles and three-dimensional structure of proteins using nuclear magnetic resonance. Some of my work in finance/business informatics is relevant to learning the business workflow from audit trial data provided by complex IT systems, identifying micro- and macro-economic factors that determine changes and dynamics in relevant business related variables, and discovering patterns in high-frequency financial data.