Introducing an integral optimised warping (IOW) approach for achieving swift alignment of drifted chromatographic peaks: an optimisation of the correlation optimised warping (COW) technique

Abstract

The drift in the position of chromatographic peaks is an unavoidable problem that needs to be taken care of before subjecting them to any data analysis work packages. The correction is essential because each of these data analysis work packages compares the chromatograms at each retention time point. Thus, it is important that a peak if present in all chromatograms must appear at the same retention time point. Correlation optimised warping (COW) has found its place in various data analysis work packages as a pre-processing technique to correct the drift in the chromatographic peak position. However, it has a disadvantage that computationally it is demanding, laborious and time consuming; as a result it becomes the bottleneck for any work package. The present work addresses this issue by introducing a novel approach baptised as integral optimised warping (IOW) that makes the correction of the drift in the peak position a simple and swift process. IOW achieves this by reducing the ambiguity in selecting the reference chromatogram and synthesising the reference chromatogram mathematically using a Gaussian function. COW optimises the combination of segment length (m) and slack (t) parameters by maximizing the Pearson correlation (ρ) coefficient between the reference and the warped chromatogram. COW calculates the ρ values for every possible combination of m and t making it a computationally challenging task. The IOW algorithm reduces the computational time that is required to optimise m and t by a significant amount by introducing mathematical criteria that allow evaluation of the integral overlap between the reference and the warped chromatogram only if the overlap occurs in a specified region around the reference peak. The IOW algorithm also reduces the computational time significantly by allowing simultaneous alignment of a specific peak in all the chromatograms. The utility of IOW is successfully demonstrated using simulated as well as real life chromatograms.