Nonparametric Models for Proteomic Peak Identification, Quantification, and Classification

• Activity Number: 393
• Type: Invited
• Date/Time: Wednesday, August 9, 2006 : 10:30 AM to 12:20 PM
• Sponsor: WNAR
• Author(s): Merlise Clyde*+ and Leanna House and Robert Wolpert
• Companies: Duke University and Duke University and Duke University
• Keywords: Bayesian ; Levy process ; kernel regression

Abstract

We present model-based inference for proteomic peak identification, quantification, and classification from mass spectroscopy, focusing on nonparametric Bayesian kernel models. Using experimental data generated from MALDI-TOF mass spectroscopy (Matrix Assisted Laser Desorption Ionization Time of Flight), we model observed intensities in the spectra with a nonparametric model for the expected intensity as a function of time-of-flight. In particular, we express the unknown function as a sum of kernel functions, which provide a natural choice of basis functions for modeling peaks. We give interpretations of model parameters in the context of the problem and how to place priors on the unknown function using Levy random field priors. Extensions of the model to multiple spectra will be presented.