TLQ Method: A Tandem Mixing-Linear Deduction-Quantile Weighted Median Framework for Robust Differential Metabolite Screening in LCMS-Based Untargeted Metabolomics
Year: 2026
Type: preprint
Abstract: Untargeted metabolomics has long been constrained by three interrelated factors: inter-individual variability, difficulty in detecting low-abundance metabolites, and matrix effects. Existing data-processing strategies tend to compensate for systematic variations through algorithmic corrections after data acquisition, but their effectiveness is limited when biological variation is substantial or signals approach detection limits. To address these issues, we propose the TLQ method. Its first module, Tandem Mixing, suppresses matrix effects and enhances signals of low-abundance metabolites through equal‑volume mixing of paired control and treatment samples; its second module, Linear Deduction, strips background through linear subtraction under minimized interference, thereby recovering the true treatment signal; its third module, Quantile Quality‑Weighted Median, performs robust screening on fold‑change data using a chromatographic peak quality-driven weighted median, and further evaluates the reliability of the fold-change direction with Bootstrap confidence intervals. The three modules can be flexibly deployed—the paired mode is optimal when combined, while unpaired scenarios can adopt the A_M mode, and the third module can be independently applied to existing data to improve screening robustness. A validation experiment based on an LPS-induced mouse inflammation model is currently ongoing. Overall, the TLQ method provides an alternative route for differential metabolite screening that compresses inter-individual variability, low-abundance detection difficulty, and matrix effects at the experimental design stage.
Source: ChemRxiv
Authors: Teng Teng, Yuchun Lu, Tonglei Qiao, Feng Liu, Chun-E Liu
Institution: China Agricultural University
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7208771260