**3. Results**

Figure 1 shows the chemical moieties, related abbreviations and the adopted labelling scheme; Figure 2 reports the GC-FID profile referring to the CSO extracted in our laboratory and Figure 3 represents the 13C-NMR profile of EVOO and CSO in the unsaturated region (127–131 ppm) along with the relative assignment witnessing the presence of the *cis*-vaccenic ester. As easily foreseeable, other NMR spectral regions also clearly showed *cis*-vaccenic resonances; however, a total assignment of 18 13C carbon atoms was challenged by the many overlaps. Previous pioneering studies pointed out the challenging quantitative decoding of the mono-unsaturated fatty esters mixture in EVOOs [23]. Specifically, other minor mono-unsaturated fatty esters (MUFE) were taken into account; beyond the oleic (O) are also considered cis-vaccenic (V), eicosenoic (E) and palmitoleic (PO) [24,25]. On the other hand, data coming from known EVOOs compositions, limit the quantitative contribution of E and PO below 1% [9] and it is consistently witnessed by the lack of defined resonances in the regions where these esters should not have overlap with other similar constructs. The Multiple Assignment Recovered Analysis (MARA-NMR) takes advantage of any spectral section also overcoming the overlap issues hampering, so far, the independent quantification of mono-unsaturated fatty esters. Specifically, in this case, MARA-NMR processing definitely led to the detection and quantification of the V esters (consistently all over the recorded spectral span). Among the 20 variables feed out from MARA-NMR whose code is reported in Table 1, we herein have restricted our considerations to the most meaningful 11 variables reported in Table 2 along with the relative standard deviation.

With respect to the other studies [24,25] the new information remarkably smooths discrepancies between 1H and 13C-NMR as the mono-unsaturated fatty esters contribution in 1H-NMR matches the contribution of O and V esters, which actually should be also somewhat enhanced by the minor PO and E esters' contribution. Because of the tricky GC-FID resolution between V and O, also referred to in the European regulation (which suggests to report the whole V+O contribution), it is not always possible to compare GC and NMR data. However, the new available data, display the best fitting so far obtainable (Figure 4) concerning the measurements of mono-unsaturated (O + V + PO), saturated (P + S), di-unsaturated (L) and tri-unsaturated (Ln) fatty esters. The average V contribution is around 3% and it is consistent with previous NMR [23] and GC-FID [26] analyses; on the other hand, we think that MARA-NMR is the most versatile method suitable for serial processing of several samples and data. We think that this remarkable parameter in EVOOs cannot be ignored, since it is not constant by shifting from sample to sample, therefore it could assess specific features of different food products. The V quantification is not a marker for this study according to Table 2; however, it will trigger many important statistical considerations.

**Figure 1.** Chemical scheme of the fatty esters commonly found in olive oils with relative abbreviation. Usually these acyl residues are esters of the glycerol moiety. The labelling scheme of carbon atoms is adopted in this paper for assignments and discussion.

This enlightened an important piece of information concerning the *cis*-vaccenic ester as main compound in CSO but also as relevant ester contributing to the EVOO mixture. This last element was incredibly ignored so far. Table S1 (Supplementary Materials) reports the extended panel of 20 quantitative variables considered in the study for 33 samples (see details in Supplementary Materials). These values are obtained by MARA-NMR—a post-processing algorithm working over the two experiments A and B type.

**Figure 2.** Expanded region of interest in the gas-chromatographic flame ionization detection (GC-FID) profile for *Capparis spinosa* oil; oleic (O) and *cis*-vaccenic (V) methyl esters are resolved for the quantification. In the case of extra-virgin oil the O peak is around 20 times more than V. Other labelled signals are linolenic (Ln), linoleic (L) and stearic (S) esters

**Figure 3.** 13C-NMR profiles for olive oil (EVOO) in gray and capparis seed oil (CSO) in black. All the assignments for oleic (O), linoleic (L) linolenic (Ln) and *cis*-vaccenic (V), with the number representing carbon atom position respect to the 1 carboxyl position, are pretty known and coherent with quantitative and literature data.

**Figure 4.** Comparison between MARA-NMR and GC-FID measured quantitative parameters referred to: (**A**) mono-unsaturated (MUFA), (**B**) saturated (SFA), (**C**) Linoleic (L) and (**D**) Linolenic (Ln) esters in relative percent ratio.
