HeuSMA: A Multigradient LC-MS Strategy for Improving Peak Identification in Untargeted Metabolomics

Abstract

Metabolomics, which involves the comprehensive analysis of small molecules within biological systems, plays a crucial role in elucidating the biochemical underpinnings of physiological processes and disease conditions. However, current coverage of the metabolome remains limited. In this study, we present a heuristic strategy for untargeted metabolomics analysis (HeuSMA) based on multiple chromatographic gradients to enhance the metabolome coverage in untargeted metabolomics. This strategy involves performing LC-MS analysis under multiple gradient conditions on a given sample (e.g., a pooled sample or a quality control sample) to obtain a comprehensive metabolomics data set, followed by constructing a heuristic peak list using a retention index system. Guided by this list, heuristic peak picking in quantitative metabolomics data is achieved. The benchmarking and validation results demonstrate that HeuSMA outperforms existing tools (such as MS-DIAL and MZmine) in terms of metabolite coverage and peak identification accuracy. Additionally, HeuSMA improves the accessibility of MS/MS data, thereby facilitating the metabolite annotation. The effectiveness of the HeuSMA strategy was further demonstrated through its application in serum metabolomics analysis of human hepatocellular carcinoma (HCC). To facilitate the adoption of the HeuSMA strategy, we also developed two user-friendly graphical interface software solutions (HPLG and HP), which automate the analysis process, enabling researchers to efficiently manage data and derive meaningful conclusions (https://github.com/Lacterd/HeuSMA).