Statistical analysis of matrix-assisted laser desorption/ionization Fourier transform ion cyclotron resonance mass spectrometry data with applications to cancer biomarker detection

Motivation. The development of better tests to detect cancer in its earliest stages is one of the most sought-after goals in medicine. Especially important are minimally invasive tests that require only blood or urine samples. By profiling oligosaccharides cleaved from glycosylated proteins shed by tumor cells into the blood stream, we hope to determine glycan profiles that will help identify cancer patients using a simple blood test.;The data used in this dissertation were generated using matrix-assisted laser desorption/ionization Fourier transform ion cyclotron resonance mass spectrometry (MALDI FT-ICR MS). We have developed novel methods for analyzing this type of mass spectrometry data--specifically, for peak detection and quantification--and applied them to eight data sets from three different types of cancer (breast, ovarian, and prostate).;Methods. (1) We model the noise part of a typical MALDI FT-ICR spectrum using techniques from time series analysis. (2) We propose a new method of baseline estimation that accounts for the unique properties of MALDI FT-ICR MS but can be easily adapted to technologies with other properties.;Results. (1) The noise part of a MALDI FT-ICR spectrum can be accurately modeled as a simple MA(6) time series with innovations given by a generalized gamma distribution with varying scale parameter but constant shape parameter and exponent. This enables us to give a reasonable criterion for classifying a peak found in an actual spectrum as either noise or signal. (2) The proposed baseline algorithm is computationally efficient and robust to the type of one-sided signal that occurs in MALDI FT-ICR spectra. (3) The techniques we have developed appear to be effective in the analysis of MALDI FT-ICR MS data. We found significant differences between control and cancer groups in seven of the eight data sets, including two structurally-related compounds that were found to be significantly different between control and cancer groups in all three types of cancer studied.;Availability. The software used to perform the analysis described in this dissertation is available in the form of an R package called FTICRMS, version 0.7, either from the Comprehensive R Archive Network (http://www.r-project.org/) or from the author.