Alignment of gas chromatography–mass spectrometry data by landmark selection from complex chemical mixtures

Analyzing the response of analytical sensors to complex chemical mixtures is difficult, as signals may overlap and may vary between experiments. The analysis and comparison of data files depends on their repeatability, and misaligned signals will not be directly comparable. We present a method for aligning chromatographic data by the selection of landmarks, the matching of these landmarks across chromatograms, the calculation of a functional approximation describing the amount of shift of the landmarks between files using cubic spline interpolation, and the subsequent shifting of the data by the application of this functional approximation so that the two files are aligned. We demonstrate the effect of this alignment on subsequent principal component analysis.