The workflow diagram describing the experimental design adopted for the untargeted metabolomic profiling of walnut seed cots (pellicle) is reported in ( Figure 1 ) showing the following steps: sample collection, sample storage, cracking and replicate composition, pellicle isolation and preparation, untargeted metabolomics.

All analysis and visualizations were performed in GraphPad PRISM v10.1.1 (GraphPad Software, La Jolla, CA).

The curated annotations were used to explore the composition of major biochemical classes and subclasses within the untargeted metabolome of the walnut pellicle. Summarization of the data included counting feature membership in various biochemical classes, subclasses, and structurally related features within a class or subclass. Each biochemical class or subclass was represented as a relative percentage over the total number of the identified features in the dataset or a specific class or subclass.

For each dataset (representing independent analytical platforms), the relative intensity of each metabolite was expressed as percentages of the total intensity for their respective platform. The total un-normalized intensity was calculated using all of the data per platform, including unidentified features ( Supplementary Table 7 ). The total intensity for each sample was determined, and then the average of these total intensities (n = 12) was calculated as the global average intensity. The average intensity across all samples (n = 12) was calculated for every metabolite and then divided by the global average intensity. These results were multiplied by 100 to express them as percentages. To understand the relative intensity of multiple metabolites to the other members within their family, the total dataset was filtered down to just members of the oxylipins, NAEs, and DA-FAs. The same calculations described above were performed to determine the percentage intensity of each query metabolite among the narrower groups.

Concerning the targeted panel, Welch’s t-test (p < 0.05) was used to determine statistical differences in the bioactive lipid concentrations between the two waste byproducts ( Supplementary Table 5 ). No corrections for multiple comparisons were employed due to the small number of technical replicates for each class.

The raw data of the untargeted metabolomic profile are given as peak heights for the quantification ion (mz value) at the specific retention time (rt value) ( Supplementary Tables 1 – 3 ). Concentrations of the detected oxylipins and endocannabinoids results of the targeted investigation are reported in Supplementary Table 5 .

Exploration of the global metabolic profile of the walnut pellicle was enabled by a triple-platform analytical pipeline, as described in the Materials and Methods section. A total of 7403 features were detected, with 1595 features identified with Platform 1 (“Complex Lipid Analysis”), 5274 with Platform 2 (“Polyphenol and Fatty Acid Analysis”), and 534 with Platform 3 (“GC-MS Analysis for primary Metabolism”). From this total number of detected features, 565 unique metabolites (216, 221, and 156 from Platforms 1, 2, and 3) were successfully identified.

Annotations for each metabolite, including biochemical pathway membership and classification by molecular structure, were used to summarize the metabolomic profile of the pellicle. As shown in Figure 3 , 56.28% (318) of the identified metabolites were classified as lipids, followed by carbohydrates at 17.17% (97), phenolics at 15.04% (85), peptides and amino acids at 5.66% (32), alkaloids at 1.77% (10) and lastly nucleic acids at 1.24% (7). Seed coat metabolomic studies have generally been interested in the phenolics fraction due to their role in barrier function, contributing antioxidant and antimicrobial properties. This has also largely been the case for walnuts, with pellicle (and shell and hull) metabolomics focused primarily on the diversity of phenolics contributing to antioxidant capacity. To our knowledge, this investigation is the first deep metabolomic profiling of the walnut pellicle, revealing a vast repertoire of diverse lipids and lipid-derived compounds.

The lipid fraction accounted for a total of 318 unique metabolites, which were identified as lipid molecules using their biochemical class annotations. The footprint of each major lipid class, represented as metabolite counts and relative percentage of the overall lipids, can be seen in Figure 4A . These major classes encompass glycerolipids, fatty acids (and fatty acyls), phospholipids, terpenoids, and sphingolipids.

The glycerolipid (GL) class was the most abundant of the lipid classes, accounting for 40.57% (129) of the total lipid features. This class includes triacylglycerols (TG), diacylglycerols (DG), and monoacylglycerols (MAG), which were represented at 84.50% (109), 10.08% (13), and 5.43% (7) of the total GL class, respectively ( Figure 4B ). Within the DGs, multiple species of glycolipids were detected and comprised 3.1% and 1.26% of the GLs and total lipid features, respectively. Of particular note are glycerolipids containing a glucuronic acid moiety, specifically two diacylglycerol glucuronide molecules named DGGA 34:2 (16:0/18:2) and DGGA 36:3 (18:1/18:2). This peculiar subclass of GLs has, to the best of our knowledge, never been reported in the walnut pellicle.

Fatty acids, belonging to the “Fatty Acids and Acyls” class, were the second most frequently identified group in the walnut seed coat, representing 21.70% (69) of the total lipid fraction expressed in counts ( Figure 4A ). A rich and highly diverse array of free fatty acids (FFAs) were identified, accounting for 63.77% (44) of the total fatty acids ( Figure 4C ). FFAs diversity was explored by grouping metabolites according to their structural characteristics, including carbon chain length, chemical modification of the carbon chain, and degree of unsaturation. The majority of FFAs, up to 70.46% (31) of the total fatty acid counts, were long-chain fatty acids, with chain lengths between 13–21 carbon atoms, followed by medium chain 11.36% (5) with chain lengths of 6–12 carbon atoms, and very long chain fatty acids 18.18% (8) with chain lengths ≥ 22 carbon atoms ( Figure 5C ). The distribution of unsaturation degree of the carbon chain in the FFAs population is shown in Figure 5B ; 40.91% (18), 27.27% (12), and 31.82% (14) accounted for by polyunsaturated FFAs (PUFAs), monounsaturated FFAs (MUFAs), and saturated FFAs (SFAs) respectively, including both oxygenated and non-oxygenated FFA ( Figure 5B ). Notably, half of the FFA, 50% (22), possessed oxygenated carbon chains. These included mono- and poly-hydroxy FFAs and mono-hydroxyperoxy and epoxy FFAs ( Figure 5D ). Typical non-oxygenated long-chain monosaturated fatty acids (LC-MUFAs) and long-chain polyunsaturated fatty acids (LC-PUFAs), which include oleic, linoleic, and linolenic acids, accounted each for just 9% (4) of the total FFAs; similarly, the low count was also found for simple long-chain saturated fatty acids (LC-SFAs) at 4.55% (2) of the total FFAs identified ( Figure 5A ). In contrast, very long-chain fatty acids (VLC-FAs) counted up to the 18.18% (8) of the total FFAs.

This larger class of fatty acids also encompassed more unusual fatty acids belonging to the classes of dicarboxylic fatty acids (DC-FAs) at 13.04% (9), fatty acid amides (FAAs) at 11.59% (8), lactones at 5.79% (4), fatty acid esters (FAEs) at 2.90% (2), and lastly fatty alcohols (FOHs) and jasmonates at 1.45% (1) ( Figure 4C ). To our knowledge, this is the first time these subclasses of fatty acids were detected in walnut seed coat, in particularly the DC-FAs and lactones, along with a substantial percentage of oxylipins, in walnut-derived tissue.

Phospholipids (PLs), terpenoids/isoprenoids, and sphingolipids (SLs) represented 16.35% (52), 11.01% (35), and 10.38% (33) of the total number of lipid features, respectively ( Figure 4A ). Concerning the PLs, phosphatidylcholines (PCs) were the most represented subclass, with 57.69% (30) of the total PL population. PCs were followed by phosphatidylinositols (PIs) at 17.31% (9), phosphatidylethanolamines (PEs) at 15.38% (8), phosphatidic acids (PAs) at 7.69% (4), and finally phosphatidylethanols (Peths) at 1.92% (1) ( Figure 4D ). A noteworthy subclass of PLs, the lysophospholipids (lyso-PLs), were also identified in this study primarily contributed by the PCs class; in total five Lyso-PCs were identified counting for the 16.67% of the total PC features and 9.67% of the total PL population ( Supplementary Table 4 ). Concerning the class of terpenoids/isoprenoids, membership was mainly represented by tetraterpenoids which reached up to 42.86% (15), followed by monoterpenoids at 22.86% (8), diterpenoids at 14.29% (5), sesquiterpenoids at 11.43% (4), and sterols at 8.57% (3) ( Figure 4E ). Lastly, the SL class was accounted for mainly by simple sphingolipids (sphingoid bases and ceramides) comprising 69.70% (23) of the total SL population. In comparison, complex sphingolipids (sphingomyelins and glycosphingolipids) accounted for the remaining 30.30% (10) of the total SL features ( Figure 4F ).

As a general conclusion, the qualitative compositional lipid profile of the walnut pellicle followed the order GL > FA > PL > Terpenoids/Isoprenoids > SL. GL and FA were the lipid classes most abundant in feature counts, with the TG and FFA subclasses contributing the highest counts of molecular species, respectively. Above all, the fatty acid profile of the walnut seed coat showed more insightful peculiarities compared to the commonly detected and extensively characterized fatty acids, like linoleic and linolenic acids, in the whole walnut kernel.

The analytical pipeline unveiled the detection of oxygenated fatty acids, i.e., oxylipins, up to 43.47% (30) of the total fatty acids. As shown in Figure 6A , most oxylipins were detected in the FFA class, followed by the lactones, DC-FAs and jasmonates classes. Compositionally, oxygenated FFAs accounted for 50% of the total FFA features ( Figure 5D ). As mentioned in the previous section, the types of oxygenation forms included mono-, di-, poly-hydroxylation, mono-hydroperoxidation, and mono-epoxidation ( Figure 6B ). To summarize, the majority of detected FFA oxylipins had long carbon chains, one or more unsaturation’s, and one or more oxygenations ( Figures 6E–G ). Looking at the combinations of these structural hallmarks, the majority of identified FFA oxylipins were long-chain polyhydroxy monounsaturated fatty acids (LC-PH-MUFA), followed by long-chain monohydroxy polyunsaturated fatty acids (LC-MH-PUFA), and long-chain mono-hydroperoxy polyunsaturated fatty acids (LC-MHp-PUFA). The list of metabolites belonging to the LC-PH-MUFA subclass is reported in Table 1 , along with each metabolite’s relative abundance (%) to the global normalized intensities and molecular structures ( Figure 6D ).

The second major subclass of fatty acids detected in this study were the DC-FAs ( Figure 4C ), which were also structurally classified based on chain length, degree of saturation, and chemical modification. DC-FAs were detected mainly as medium chains with similar frequencies of mono- and poly-unsaturation’s ( Figures 7A, C, D ). While the majority were non-oxygenated, several oxylipins of DC-FAs were detected, making up 37% of the total DC-FAs features ( Figure 7E ). These results are seen in Table 2 , which reports a list of the identified DC-FAs in the walnut pellicle along with their relative abundance (%) and molecular structures ( Figure 7B ). Interestingly, methyl esters of DC-FAs were also identified and classified under the fatty acid ester (FAE) class.

The third most abundant subclass of fatty acids was the fatty acid amides (FAAs). FAAs were classified as fatty acid primary amides (FAPAs), N-acylethanolamines (NAEs), and ethylamides (FA-EtAs), which accounted for 50% (4), 37% (3), and 12.50% (1) of the total FAA profile, respectively ( Figure 8A ). The identified FAAs possessed chain lengths of C16 and C18, including palmitoyl ethanolamide, palmitamide, stearamide, oleoylethanolamide, and oleamide. Erucamide was the only FAA with a C22 carbon chain. The list of these species and their relative abundances (%) are reported in Table 3 , with their molecular structures shown in Figure 8B .

After their detection in the untargeted metabolomics analysis, oxylipins and N-acylethanolamines (NAEs) were quantified in two industrial waste streams from handling processes of walnuts in California, named “Sorting Meal Room (SRM)” and “Blower Fluff (BF)”. While not the only walnut processing waste streams, SRM and BL were selected for being enriched in pellicle and generally considered as abundant, renewable, and low-value agricultural waste. These wastes were of particular interest for potential re-utilization and valorization into high-value products due to their array of potent bioactive lipids relevant to plant and human health.

Oxylipins derived from free fatty acids and NAEs, also called Acyl-ethanolamides (Acyl-EAs), were quantified in SRM and BF wastes along with ketones, PUFAs, Acyl-Amino Acids (Acyl-AAs), and prostaglandins (PGs). The 74-member panel of metabolites was identified and quantified by targeted metabolomic. As for percentages, oxylipins composed 74.68% of the panel, followed by Acyl-EAs (11.39%), PUFAs (6.33%), Acyl-AAs, and PGs (3.8%). Within the oxylipins group, metabolites were distinguished by oxygenation type ( Figure 9A ): including epoxides of fatty acids (21.52%), monohydroxy fatty acids (R-OH) (20.25%), vicinal diols (vic-Diols) (21.52%), ketones (R=O) (7.59%), triol fatty acids (2.53%), and hydroperoxy fatty acids (R-OOH) (1.27%).

Concentrations of monohydroxy fatty acids, including 13-, 9-, 12-, and 10-HODE, were the highest along with triols, like 9,12,13-TriHOME ( Figures 9B, E ), in both waste streams, reaching concentrations of 80–100 μM. Comparing the two waste streams, BF was more enriched in vicinal diols like 15,16-, 12,13-, and 9,10-DiHODE ( Figure 9D ). These subsets of metabolites are reported as relative abundance across all measured samples (i.e., the sum of each metabolite across all samples is set to 100%) rather than concentrations in μM. Along similar lines, 9,10-DiHO, as well as 12,13- and 9,19-DiHOME, were also more abundant in BF and had lower concentration compared to triols and R-OH ( Figure 9C ). Epoxides of fatty acids were enriched in both wastes: 9(10)EpO, 12(13)EpOME, and 9(10)EpOME were detected in concentrations of 10–35 μM ( Figure 9F ).

A rich profile of NAEs was detected and quantified ( Figure 9G ). Linoleoylethanolamide (N-C18:2n6_EA) was the most abundant NAE detected in both waste streams with an average value of 5.7 μM, followed by oleoylethanolamide (N-C18:1n9_EA) at 1.2 μM, linolenoylethanolamide (N-C18:3N3_EA) at 0.9 μM, palmitoylethanolamide (N-C16:0_EA) at 0.69 μM, and finally stearoylethanolamide (N-C18:0_EA) at 0.5 μM. The complete list of metabolites detected and quantified in both waste by-products is included in Supplementary Table 5 .

Establishing a successful process for pellicle isolation from nutmeat enabled the identification of metabolism specifically occurring in the pellicle. An untargeted metabolomic analysis enabled a comprehensive qualitative description of pellicle lipid composition. The proportion and composition of major lipid classes, expressed as percentages of total feature counts, were broadly consistent between isolated pellicles and previously studied kernels (including the nutmeat and the pellicle in most published studies), although with interesting differences.

Free fatty acid (FFA) species were also found to represent a considerable portion of the pellicle lipid population. Shown extensively in the literature, walnut kernels possess a high content of polyunsaturated fatty acids (PUFAs), especially linoleic acid, and the lowest saturated fatty acids: total fatty acids ratio and greatest content of PUFAs compared with other nuts (Vecka et al., 2019; Jardim et al., 2023). Accumulation of long-chain (LC) FFAs like linoleic, oleic, and linolenic acids has also been shown in dried walnut shells and husk tissue (Ventura et al., 2023; Herrera et al., 2020; Romano et al., 2021; Barekat et al., 2023). In walnut seed coat, LC-PUFAs, LC-MUFAs and LC-SFAs, including their oxygenated derivatives called oxylipins, accounted for 40.91% (18), 22.72% (10), and 6.82% (3), respectively of the total FFA counts (44 features) and roughly 5%, 3% and 1% of the total identified lipid features (318 features). In general, lipidomic profile of walnut kernel is more enriched, both quantitatively and qualitatively, in unsaturated fatty acids than saturated (Yan et al., 2021; Okatan et al., 2022; Jardim et al., 2023; Zhang R. et al., 2021), a trend recently shown to be conserved across several walnut genotypes (Okatan et al., 2022). Our study provides qualitative results which support previous research; indeed, similarly to kernel FFA composition, the majority of the FFAs detected in walnut seed coat were LC-PUFAs and LC-MUFAs, summing up to the 63.7% of the total FFA features’ count. More broadly, Figure 5B shows distribution of degree of unsaturation in the entire class of FFAs detected in our data, including very long-, long- and, medium-chain fatty acids, as well as oxygenated and non-oxygenated FFAs. This confirms that polyunsaturation is the most prevalent feature of FFAs class in walnut seed coat.

The majority of LC-MUFA and LC-PUFA features detected in walnut pellicle were oxygenated in the form of hydroxy or epoxy long-chain fatty acids ( Figure 5A ). While non-oxygenated LC-MUFAs and LC-PUFAs represented just the 9% each of the total FFAs, oxygenated derivatives of LC-MUFAs and LC-PUFAs (shown in Figure 5A ) reached 13.64% (6) and 31.82% (14) of the total FFAs, respectively. These results highlighted the accumulation of oxylipin derivatives of conventional fatty acids, like linoleic and linolenic acids, in walnut seedcoat, therefore suggesting that levels of oxygenated LC-MUFAs and LC-PUFAs might differ between nutmeat and seedcoat. To the best of our knowledge, no studies have been published showing LC-MUFA and/or LC-PUFA quantification specifically in walnut seed coat, as well as the quantification of their oxylipin derivatives in kernel tissues. Nevertheless, considering together the numerous lipidomic studies on the topic and our results, we could not rule out the possibility that concentrations of LC-MUFAs and LC-PUFAs would be more enriched in nutmeat than in the seed coat, where those fatty acids might be preferentially metabolized into their oxylipins derivatives. This hypothesis is valuable in light of the biological function of oxylipins in walnut seed-coat as protection against stress for the nutmeat as explained in section 4.3.1. Targeted quantitative analyses of those fatty acids in both seedcoat and nutmeat is needed to further validate this hypothesis.

Also revealed in this study was the presence of very long chain (VLC) FAs, with carbon chain lengths containing 22–30 carbon atoms, which are rare and in low abundance within the kernel (Yan et al., 2021; Jardim et al., 2023). These species accounted for 18.18% (8) of the total FFA counts in the pellicle ( Figures 5A, C ). As a minor class of the FFAs, medium chain (MC) FAs were identified at similar abundance to the kernel (Yan et al., 2021), making up 11.36% of the FFA counts ( Figures 5A, C ). Overall, this analysis revealed the pellicle FFA profile to be compositionally distinct from the reported kernel FFA profile, with the pellicle containing many uncommon metabolites to be discussed later ( Figures 4C , 6C ).

The reported walnut kernel lipid profile generally places glycerolipids (GLs) as the predominant lipid class which was mirrored in the pellicle, with triacylglycerols (TGs) being the most abundant subclass followed by diacylglycerols (DGs) and monoacylglycerols (MGs) (Yan et al., 2021; Zhang R. et al., 2021; Wang et al., 2022). Interestingly, galactolipids falling under the larger DG class were detected and consistent with prior reporting. These include digalactosyldiacylglycerol (DGDG), monogalactosyldiacylglycerol (MGDG), and sulfoquinovosyl diacylglycerol (SQDG), with DGDG and MGDG accounting for less than the 5% and 2% of the total GLs and lipid classes respectively (Yan et al., 2021; Zhang R. et al., 2021). The levels of these compounds are influenced by drying and may be accumulated as the pellicle matures and desiccates, as well as from commercial drying processes (Wang et al., 2022).

This agreement between pellicle and kernel composition was largely the case for phospholipids (PLs), sphingolipids (SLs), and terpenoids/isoprenoids. Concerning the PLs, PCs were the most accounted for subclass, followed by phosphatidylinositols (PIs), phosphatidylethanols (PEs), and phosphatidic acids (PAs) (Yan et al., 2021; Zhang R. et al., 2021; Wang et al., 2022; Jardim et al., 2023). Furthermore, high frequency of lyso-PCs, up to 24% (Yan et al., 2021) and 26% (Zhang R. et al., 2021) of the total PC population and 5% (Yan et al., 2021) and 6% (Zhang R. et al., 2021) of the total PL population, was found in the walnut kernel in other studies. These species, which arise from enzymatically-mediated PC hydrolysis by the action of phospholipases, like PLA2, can be highly dependent on genotype, tissue maturity, and abiotic stresses imposed by management practices such as drying (Wang et al., 2022) and storage conditions (Zhang R. et al., 2021). Regarding SLs, the walnut kernel and isolated pellicle are more enriched in simple sphingolipids like ceramides than complex sphingolipids like HexCer. The SL population makes up roughly 10% of the total lipid metabolites for the pellicle, similar to the case of the kernel (Yan et al., 2021; Zhang R. et al., 2021). Similar to the PLs, the increase in SLs was reported in the kernel to be an outcome of the drying process (Wang et al., 2022). As for terpenoids/isoprenoid class, phytosterols (i.e., steroidal alcohols) belonging to the group of triterpenoids were mainly found in nuts, including whole walnuts (Jardim et al., 2023) and walnut oil (Rébufa et al., 2022), while a more complex profile for other terpenoid subclasses was found for walnut leaves by VOC analysis (Valdés García et al., 2021). Unlike the kernel, the pellicle was revealed to possess a more diverse repertoire of terpenoid subclasses, including mono-, di-, tetra-, and sesquiterpenoids (in addition to the phytosterols as mentioned above). While monoterpenoids were previously identified in walnut kernels by supercritical fluid extraction coupled to GC-MS analysis, detection of further subclasses was not reported (Kessler et al., 2023). Overall, these results are broadly consistent with the published data for the walnut kernel (Huang et al., 2021; Yan et al., 2021; Zhang L. et al., 2021; Zhang R. et al., 2021; Wang et al., 2022) obtained with analytical pipeline similar to one adopted for this study.

A summary of the lipid compositional profile of walnut tissues reported in most recent publications is presented in Table 4 .

As previously mentioned, fatty acid species comprised 21.70% of the total lipids detected in the walnut pellicle, more than twice the number detected in the kernel by Yan et al. (Yan et al., 2021) and Wang et al. (Wang et al., 2022). The FFAs as a subclass constituted 63.77% of the FA profile, and similar to the kernel (Yan et al., 2021; Wang et al., 2022), were primarily LC-MUFAs and LC-PUFAs ( Figures 5A–C ). However, these groups contained many unusual members aside from the conventional FAs (i.e., C18:1 oleic, C18:2 linoleic, and C18:3 linoleic acid) in the pellicle. Half of the total FFAs (50%) were oxygenated in the form of hydroxy-, peroxy-, or epoxy-fatty acids, which fall into the umbrella term of “oxylipins” ( Figures 5D , 6B ). These FFA-derived oxylipins are also involved in the production of related molecular families, including jasmonic acid (Acosta et al., 2009) and certain γ-lactones that may contribute to the barrier function of the pellicle. Independent of the oxylipins, the second and third significant subclasses of detected FFAs were the dicarboxylic FAs (DC-FAs) and the fatty acids amides (FAAs). To our knowledge, this is the first reported detection of DC-FA in walnut tissue, as prior metabolomic studies of walnuts did not report this unique subclass of FFAs. Regarding the FAAs, of which N-acylethanolamines (NAEs) are a subgroup, detection and quantification of NAEs in the whole walnut kernel have been reported previously (De Luca et al., 2019).

Intriguingly, the same classes of bioactive lipids identified in this study show potent bioactivities in mammals (notably humans) and are pivotal modulators of various physiological processes related to health and wellbeing. As previously mentioned, bioactive lipids can have peculiar structural features, including oxygenations of the carbon chain, conjugation to other bioactive molecules (like amino acids), or unusual degree of unsaturation (Cahoon and Li-Beisson, 2020).

Considering the bioactive lipids’ relevance for human clinical outcomes, this broad group of compounds should be regarded as “essential bioactive dietary lipids”. Therefore, exploring edible dietary sources highly enriched in oxylipins, DC-FAs, and para-endocannabinoids (like NAEs and Acyl-AAs) is a promising avenue for improving the incorporation of these compounds into the human diet. Identifying these bioactive lipids in the walnut pellicle paves the way for identifying molecular features and pathways at the origin of the plethora of currently unexplained clinical outcomes (from cohort studies and randomized controlled trials) from walnut consumption. These outcomes range from reducing the risk of certain cancers, regulation of dysfunctional metabolic conditions like diabetes and obesity, improvement of cognitive function, amelioration of mood disorders and addiction behaviors, support of gut health by modulation of the gut microbiome, amelioration of male fertility, and age-related cognitive decline. While these conditions are major societal concerns, the latter are becoming increasingly important as the global population is aging and the fertility rate is declining in the industrialized countries. The consumption of functional foods to improve these various aspects of health and well-being is a significant trend among young adult consumers. For this reason, identifying bioactive lipids and their associated clinical outcomes might encourage walnut consumption in the younger population. Furthermore, the discovery of a natural source of bioactive lipids in the form of waste byproducts from walnut industrial processing will pave the way for developing novel edible supplements, nutraceuticals, or precursors for derma care products and pharmaceuticals. This discovery has significant economic impacts as rising walnut production will lead to a commensurate rise in waste by-products from the handling and processing pipelines. What have been classically low-value waste by-products may now be an opportunity for walnut industries to upcycle, encouraging sustainability and creating additional revenue streams to support industry longevity. Finally, walnut improvement programs may leverage breeding or bio-engineering approaches to enhance the production of these various bioactive lipids in the walnut pellicle or other walnut tissues to supply greater concentrations of these potent molecules for applications in human health.