Fig. 7. Evaluation of feature quantification.

Scatter plot of the log₂ peak areas of common features between asari and XCMS, on three samples in HZV029 dataset (a) and Yeast2021 dataset (b). The r values are based on Pearson correlation. c Design of the BM21 dataset, by varying mix ratios between human plasma and vegetable juice. A well quantified metabolite is expected to show good correlation between the mixing ratios and the reported peak areas, as exemplified by the feature on top (m/z 189.1232, 159 s). Asari calculates peak area differently from XCMS, resulting in higher values in Orbitrap data. d Overall quantification results in the BM21 dataset, shown as feature numbers binned by Pearson correlation coefficients between peak areas and sample mixing ratios. Source data are provided as a Source Data file.