Department of Computer Science: MSc Thesis Presentation
Date: Thursday, 19 May 2022
Detection and integration of chromatographic peaks using theoretical peak fitting
Author: Juho Kuikka
Advisor: Thiago Brito, Thermo Fisher Scientific
Supervisor: Juho Rousu
Peak detection is a fundamental part of chromatography data analysis. In liquid chromatography-mass spectrometry (LC-MS) method development, accurate peak detection is crucial to producing reliable results. Noise in the chromatogram, among other things, can make peak detection challenging. In this Master's thesis, a literature review of peak detection and integration methods is presented, as well as methods used to estimate noise in the chromatogram signal prior to peak detection. A novel framework for detecting chromatographic peaks that uses theoretical peak shapes is also introduced.
The second part of this thesis consists of a case study where the goal is to reduce the relative standard deviation of the integrated peak areas, since this parameter correlates with the method performance. A peak detection and integration framework was developed and applied to data sets produced with the LC/MS method. The framework had a couple of options for estimating noise in the chromatogram data and detecting the start and end points of the peak. The methods for detecting the start and end points of peaks is based on fitting a theoretical peak to the chromatograms. The relative standard deviations obtained with the framework are compared to the relative standard deviations obtained with a known peak detection method from the literature and to the currently used method in the analyzers.
The proposed framework performed well compared to the other methods, despite the noise in the data and the varying peak shapes. In most of the analysed data sets, the proposed framework was able to produce lower relative standard deviations of the integrated areas. It was concluded that fitting a theoretical peak improved the precision of the detection and integration of chromatographic peaks.