Multiscale peak alignment for chromatographic datasets

• Published in: Journal of Chromatography A 2012-02, Vol.1223, p.93-106
• Main Authors: Zhang, Zhi-Min, Liang, Yi-Zeng, Lu, Hong-Mei, Tan, Bin-Bin, Xu, Xiao-Na, Ferro, Miguel
• Format: Article
• Language: English
• Subjects: chemometrics; Chromatography; Cross correlation; data collection; Fast Fourier transform; herbal medicines; metabolomics; Peak alignment; quality control; Shannon information content; statistics; wavelet
• ISSN: 0021-9673
• DOI: 10.1016/j.chroma.2011.12.047

► We proposed MSPA method to align chromatograms accurately and rapidly. ► MSPA can preserve shapes of peaks (peak height and area) during alignment. ► It has excellent aligning speed because of FFT cross correlation. ► It is robust and not sensitive to noise and baseline. ► The source code of MSPA is available at http://code.google.com/p/mspa. Chromatography has been extensively applied in many fields, such as metabolomics and quality control of herbal medicines. Preprocessing, especially peak alignment, is a time-consuming task prior to the extraction of useful information from the datasets by chemometrics and statistics. To accurately and rapidly align shift peaks among one-dimensional chromatograms, multiscale peak alignment (MSPA) is presented in this research. Peaks of each chromatogram were detected based on continuous wavelet transform (CWT) and aligned against a reference chromatogram from large to small scale gradually, and the aligning procedure is accelerated by fast Fourier transform cross correlation. The presented method was compared with two widely used alignment methods on chromatographic dataset, which demonstrates that MSPA can preserve the shapes of peaks and has an excellent speed during alignment. Furthermore, MSPA method is robust and not sensitive to noise and baseline. MSPA was implemented and is available at http://code.google.com/p/mspa.