Unraveling non-target screening variability for LC-HRMS data: a chemometric comparative analysis of river water samples impacted by treated wastewater.

Non-target screening (NTS) using liquid chromatography-high-resolution mass spectrometry (LC-HRMS) is crucial for identifying organic compounds and patterns in complex environmental samples. However, discrepancies in data processing between peak-picking algorithms ("feature profile" approaches) remain a critical challenge in achieving consistent and reproducible results. An alternative approach employs multi-way chemometric methods to directly produce "component profiles," allowing efficient decomposition and evaluation of LC-HRMS datasets, often after data compression. This study compares two distinct NTS workflows-MZmine3 and regions of interest multivariate curve resolution-alternating least squares (ROIMCR)-using surface water samples from a large-scale mesocosm experiment receiving wastewater effluent over a 10-day exposure period. Samples were analyzed in both positive and negative ionization modes to assess workflow influence on subsequent multivariate analyses. ANOVA simultaneous component analysis (ASCA) quantified treatment and temporal effects, identified dynamic patterns, and prioritized relevant features, while partial least squares discriminant analysis (PLS-DA) discriminated effect-based classes and highlighted significant chemical components and features. Results demonstrated that both workflows significantly differentiated treatment and temporal effects but exhibited different characteristics. Temporal variation dominated through implementing ROIMCR (35.5-70.6% variance), whereas MZmine3 showed comparable contributions from time (20.5-31.8%) and sample type (11.6-22.8%). MZmine3 showed an increased sensitivity to treatment effects but increased susceptibility to false positives. ROIMCR provided superior consistency, reproducibility, and temporal clarity but lower treatment sensitivity. Additionally, workflow agreement diminished with more specialized analytical objectives and prioritized features through implementing multivariate chemometric approaches, highlighting the non-holistic capabilities of individual NTS workflows and the potential benefits of their complementary use.