4.4.2. Spectra Acquisition

1H-NMR experiments were performed on Bruker Avance 600 MHz equipped with a SampleJet autosampler using a noesypr1d sequence, mixing time of 100 ms, a spectral window of 12 ppm, acquisition time of 2 s, relaxing time of 3 s, 516 scans, 4 dummy scans, and T = 298 K. This sequence has become the best choice for NMR-based metabolomics studies [42] for several reasons. Firstly, the quality of water suppression is very high without the need for extensive optimization. Secondly, an increasing number of wellestablished groups utilize the sequence, reflecting its consistency [43]. Finally, the library of Chenomx used in this study to quantify metabolite concentrations is optimized for this sequence and compensates for incomplete relaxation.

#### 4.4.3. H-NMR Data Analysis

All the spectra were processed using 0.5 Hz of line-broadening followed by manual phase and baseline correction. Chenomx NMRSuite 8.5 (Chenomx Inc.) was used to quantify the concentrations of the metabolites. The spectra database in this software allows for the manual deconvolution of different signals and determines the concentration of the compounds that form the mixture. TSP was set as an internal standard at 0.5 mM.

#### *4.5. SYNHMET Method*

The starting spectrum profile for deconvolution is defined using the average concentrations of urine metabolites [16]. The chemical shifts and levels of all compounds are then varied to reproduce the profile observed in each experimental NMR spectrum. The matching between the calculated and experimental spectral profiles is never perfect. The source of this inequality can be understood by analyzing all variables contributing to the spectrum intensity at a given chemical shift (*Ik*) (Equation (1)):

$$I\_k = \sum\_{i=1}^n a\_i K\_{i,k} + \sum\_{j=1}^m b\_j \mathcal{U}\_{j,k} + \mathcal{N}\_k \tag{1}$$

where *k* is the chemical shift, *i* represents one assigned metabolite, *n* is the total number of assigned metabolites, *Ki,k* is a known factor accounting for the shape of assigned metabolites, *j* represents one unassigned metabolite, *m* is the total number of unassigned metabolites, *Uj,k* is an unknown factor considering the shape of unidentified metabolites, *ai* and *bj* are the metabolite concentrations, and *Nk* is a random factor representing the noise.

In parallel, the exact mass of each metabolite is searched in the MS dataset, creating a list of linked MS features for most compounds. The number of MS peaks associated with each metabolite varies from zero to more than twenty. Detecting more than one peak with the same exact mass turns the identification based solely on the molecular weight uncertain unless using labeled standards. In the SYNHMET method, combining the concentrations measured for a cohort of samples simultaneously by MS and NMR can solve this ambiguity in an alternative way. We considered that a certain MS-detected chromatographic peak showing the accurate mass of a metabolite can be attributed to it when it is the only one showing a significant correlation between the distributions of the MS peak intensities and NMR concentrations. The intensities of the selected peak are then converted into concentrations by multiplying them with the slope of the best fit solution. The initial spectrum profile is then adjusted, inserting the values of the peak or peaks averaged to those measured by NMR for each metabolite. Conversely, concentrations of compounds not represented by any MS feature or showing multiple or no correlations are not updated for the following phase.

During the next profiling step, all compounds' signal positions and concentrations defining the updated profile are varied to obtain the best accordance between the calculated and experimental profiles. After completion, a new correlation test is accomplished, possibly increasing the number of identified and consequently quantified metabolites. This process is iteratively repeated until no further information is added. The final matrix

contains concentrations of metabolites that are determined by a combination of MS and NMR measurements.

#### **5. Conclusions**

In conclusion, the new methodology for merging NMR and UHPLC–HRMS produced a list of 165 metabolite concentrations in urine in almost all samples, with significantly higher accuracy of identification and quantification than could be reached separately using the two techniques. In addition, its application allowed us to delineate a personalized urinary profile based on a list of compound levels covering a wide range of metabolic processes. Its expansion to more samples in the future will allow us to enlarge our knowledge of many metabolites' normal and abnormal values in human urine. Its translation into clinical practice can be of great value, such as identifying biomarkers of disease susceptibility and following the individual therapeutic outcomes [35]. These two aspects are among the main applications of metabolomics to improve the accuracy of personalized medicine.

**Supplementary Materials:** The following are available online. Table S1: List of the 164 metabolites identified and quantified by the SYNHMET approach; Figure S1: Personalized metabolic profile for subject 2852 showing the 164 metabolites identified and quantified by SYNHMET; Table S2: Comparison of the retention times observed for nine labeled standards with those assigned with the NMR–HRMS intensity correlation method in the two chromatographic conditions.

**Author Contributions:** Conceptualization, G.P. and D.O.C.; methodology, G.P. and D.O.C.; validation, G.P., G.C. and D.V.; formal analysis, G.P.; investigation, G.P. and D.O.C.; resources, E.M. and V.S.; data curation, G.P., S.L., C.M., R.S., A.S., F.M., R.V. and L.O.; writing—original draft preparation, D.O.C.; writing—review and editing, G.P., E.M. and D.O.C.; funding acquisition, V.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Collezione Nazionale di Composti Chimici e Centro Screening (CNCCS) Consortium, Grant: Project B, Sp 2, WP2, 2019 "Metabolomics Studies".

**Institutional Review Board Statement:** All the studies carried out on patients' samples were conducted according to the guidelines of the Declaration of Helsinki and approved by the Institutional Ethical Committee (IRCCS Ospedale San Raffaele, Milan, protocol URINEBIOMAR, approval date 14 July 2016).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

**Sample Availability:** Samples are not available from the authors.

## **References**

