2.2.4. Final Results

Table 2 summarizes the results obtained after two deconvolution rounds for the eleven metabolites clustered into the region of 2.47–2.37 ppm. Nine out of eleven metabolites could be unambiguously linked to one or more MS hits. Due to the different chromatographic conditions and polarities used, a single compound could be represented by more than one peak, such as for 4-pyridoxate, carnitine, and succinate. The final concentrations were calculated for all these nine metabolites by averaging the values obtained from the MS and NMR measurements, which showed a significantly improved correlation after the second round. This increase in the R2 value was a natural consequence of using the deconvolution process assisted by UHPLC–HRMS intensities, but its final value was still a measurement of the degree of agreement between these two datasets measured orthogonally. For 3 hydroxybutyrate and trans-4-hydroxy-L-proline, we were not able to identify any MS hits correlating with the NMR concentrations. In the first case, the final concentrations were still measured by NMR because we judged their values to be sufficiently reliable. On the contrary, the concentrations of trans-4-hydroxy-L-proline were not included in our final results. This compound has signals with high multiplicity, which divides the intensity into multiple components and does not present any resonance in a non-crowded region of the spectrum. These two facts make the NMR measurements inaccurate, hindering our ability to identify the correct MS hit(s) among the eleven showing its exact mass.

This procedure was repeated for all the regions into which the NMR spectrum was divided. As a result, we were able to quantify 165 metabolites out of the 180 initially considered. Of this total, twelve were quantified using only NMR data. Our final concentration matrix contained only 48 missing values, representing 0.6% of the total. These concentrations were normalized by converting them into μM/mM of creatinine, and they were compared with those of the literature (Table S1). The excellent agreement between the retention times of nine labeled standards co-injected with the samples with those obtained by the HRMS-NMR correlation method (Table S2) further sustains the assignments of the metabolites listed in Table S1.

The set of metabolites quantified cover a wide range of biochemical markers, including amino acids and their metabolism, markers of vitamins, dysbiosis, diet and toxin exposure, carbohydrates and their metabolism, energy, fatty acid/lipid, and glycine/serine metabolism, and ketone bodies. In this way, it is possible to cover some of the central metabolic pathways for metabolomics studies to discover biomarkers related to pathological states and individual profiling.


**Table 2.** Identified and quantified metabolites belonging to the example region between 2.47 and 2.37 ppm. The table shows for each compound the diagnostic chemical shifts for their identification and quantification, the number of chromatographic peaks detected in the mass dataset, the chromatographic peak(s) identified by the correlation with NMR data, and the value of Pearson's correlation coefficient in the first and second deconvolution step.

<sup>1</sup> Number of chromatographic peaks found in the MS dataset. Due to the different chromatographic conditions and polarities used, a single compound could be represented by more than one peak. <sup>2</sup> Chromatographic peak/s identified by correlation. Pearson correlation after <sup>3</sup> first round and <sup>4</sup> second round.
