Title: "Inferential approaches to peptide sequence identification and quantitative protein profiling in mass spectrometry-based proteomics. "

Speaker: Olga Vitek, Departments of Statistics and Computer Science, Purdue University
Place: Krannert (KRAN) G016; September 26, 2006, Tuesday, 4:30pm

Abstract

Tandem mass spectra (MS/MS) are now routinely used for identification of components in complex biological mixtures. Typically, identifications are performed by matching the observed MS/MS spectra to theoretical spectra of peptides in a database. The approach presents difficulties of statistical nature in choosing a scoring function of the match, and in controlling the false discovery rate among the identified spectra. I will discuss two scoring functions, namely the Sequest score and the PeptideProphet score, and the properties of several alternative methods controlling the false discovery rate.

Label-free liquid chromatography coupled with mass spectrometry (LC-MS) experiments are increasingly popular for profiling changes in peptide and protein abundance across samples of different biological nature, and for discovery of molecular biomarkers of disease. However, the experiments provide a sparse information regarding sequence identities of LC-MS features, and contain a considerable amount of noise. In the second part of the talk I will discuss how linear mixed models can be used to take advantage of the correlation structure of the data to simultaneously improve the identification of LC-MS features and the detection of differential abundance.




Click here for a full schedule of BIOINFORMATICS SEMINARS, past and present.