Batch alignment via retention orders for preprocessing large-scale multi-batch LC-MS experiments.

Area of Science:

• Biochemistry
• Analytical Chemistry
• Bioinformatics

Background:

• Accurate preprocessing of liquid chromatography-mass spectrometry (LC-MS) data is vital for reliable analysis.
• Multi-batch LC-MS experiments face challenges with mass-to-charge ratio and retention time shifts, complicating parameter selection and data integration.
• Individual batch preprocessing can mitigate these issues but necessitates robust alignment and combination strategies.

Purpose of the Study:

• To develop and evaluate methods for aligning and combining individually preprocessed batches in multi-batch LC-MS experiments.
• To address the challenge of minimizing information loss during the integration of large-scale, multi-batch LC-MS datasets.
• To investigate the frequency of retention order swaps in untargeted LC-MS data.

Main Methods:

• Development of two novel algorithms: kmersAlignment and rtcorrectedAlignment.
• Testing of the developed methods on six simulated and six real-world multi-batch LC-MS datasets.
• Estimation of peak insertion, deletion, and swap probabilities between batches in authentic datasets.

Main Results:

• The proposed alignment methods effectively combine individually preprocessed batches from multi-batch LC-MS experiments.
• Analysis of real datasets revealed that retention order swaps are common occurrences in untargeted LC-MS data.
• The methods were validated on both simulated and authentic LC-MS datasets, demonstrating their applicability.

Conclusions:

• The presented alignment and combination strategies offer a solution for integrating data from individually preprocessed batches in multi-batch LC-MS studies.
• Understanding the prevalence of retention order swaps is crucial for accurate data interpretation in untargeted metabolomics.
• The developed algorithms provide a valuable tool for researchers working with large-scale LC-MS data.