Distinct chemical signatures and quality markers for the discrimination of civet coffee: A metabolomic analysis

Assinaturas químicas distintas e marcadores de qualidade para a discriminação do café de civeta: Uma análise metabolômica

ABSTRACT

Civet coffee is a high-value product distinguished by biochemical modifications occurring during passage through the civet gastrointestinal tract. This study aimed to characterize the metabolomic differences between green civet coffee and unprocessed Coffea arabica beans from Vietnam using an untargeted GC-MS-based metabolomics approach for the identification of potential metabolite markers associated with civet coffee authentication. Multivariate analyses, including PCA, PLS-DA, and OPLS-DA, demonstrated a clear separation between the two sample groups, confirming distinct metabolomic profiles. Notably, most detected metabolites, particularly sugars, organic acids, amino acids, and fatty acids, were downregulated in civet coffee, whereas mannitol was consistently upregulated. In total, 23 metabolites were identified as potential discriminative markers contributing to sample classification. Correlation network and pathway enrichment analyses further revealed coordinated alterations in carbon metabolism and amino acid biosynthesis, suggesting that digestive fermentation may induce a system-wide metabolic shift. Overall, these findings provide preliminary evidence supporting the use of metabolite-based approaches for civet coffee authentication and quality control.

Index terms:
Untargeted profiling; food authentication; GC-MS; Coffea arabica

RESUMO

O café civeta é um produto de alto valor agregado, caracterizado por modificações bioquímicas que ocorrem durante a passagem dos grãos pelo trato gastrointestinal da civeta. Este estudo teve como objetivo caracterizar as diferenças metabolômicas entre grãos verdes de café civeta e grãos não processados de Coffea arabica provenientes do Vietnã, utilizando uma abordagem metabolômica não direcionada baseada em GC-MS para identificar de potenciais marcadores metabólicos associados à autenticação do café civeta. As análises multivariadas, incluindo PCA, PLS-DA e OPLS-DA, demonstraram uma clara separação entre os dois grupos de amostras, confirmando perfis metabolômicos distintos. Notavelmente, a maioria dos metabólitos detectados, especialmente açúcares, ácidos orgânicos, aminoácidos e ácidos graxos, apresentou regulação negativa no café civeta, enquanto o manitol foi consistentemente regulado positivamente. Ao todo, 23 metabólitos foram identificados como potenciais marcadores discriminativos que contribuíram para a classificação das amostras. As análises de rede de correlação e enriquecimento de vias metabólicas revelaram ainda alterações coordenadas no metabolismo do carbono e na biossíntese de aminoácidos, sugerindo que a fermentação digestiva pode induzir uma mudança metabólica sistêmica. De modo geral, esses resultados fornecem evidências preliminares que sustentam o uso de abordagens baseadas em metabólitos para a autenticação e o controle de qualidade do café civeta.

Termos para Indexação:
Perfilamento não direcionado; autenticação de alimentos; GC-MS; Coffea arabica

Introduction

Coffee quality is strongly influenced by the chemical composition of coffee beans. Among specialty products, civet coffee is one of the most distinctive and expensive types, produced after coffee cherries pass through the gastrointestinal tract of the Asian palm civet (Paradoxurus hermaphroditus). During this process, exposure to digestive enzymes and microbial activity alters the chemical composition of the beans, contributing to their distinctive sensory properties (Haile & Kang, 2019; Marcone, 2004).

Digestive fermentation may induce metabolic changes affecting proteins, carbohydrates, organic acids, and secondary metabolites in coffee beans (Muzaifa et al., 2020). As a result, civet coffee exhibits a metabolite profile distinct from conventional coffee, with reported differences in organic acids, amino acids, lipids, and alkaloids (Chan & Garcia, 2011; Farah et al., 2005; Ginz & Engelhardt, 2001). For example, NMR-based metabolomic studies reported elevated levels of compounds such as alanine, citrate, lactate, malate, and trigonelline in civet coffee (Farag et al., 2023; Gigl et al., 2021; Jumhawan et al., 2013).

Metabolomics provides a comprehensive approach for profiling food composition using platforms such as NMR, gas chromatography-mass spectrometry (GC-MS), and liquid chromatography-mass spectrometry (LC-MS) combined with multivariate analysis (Erban et al., 2019; Selamat, Rozani & Murugesu, 2021; Zhong et al., 2022). Previous studies have applied these techniques to differentiate civet and conventional coffee, identifying markers such as citric acid, malic acid, and the inositol/pyroglutamic acid ratio (Farag et al., 2023; Zayed et al., 2022). In addition, distinct metabolite signatures in civet coffee have been revealed across analytical platforms such as NMR (Farag et al., 2023; Gigl et al., 2021; Jumhawan et al., 2013), LC-MS (Farag et al., 2023; Gigl et al., 2021), and GC-MS (Garrett et al., 2014). However, most studies have focused on roasted coffee, where thermal processes such as Maillard reactions, caramelization, and lipid oxidation can obscure fermentation-induced differences (Ollennu-Chuasam et al., 2025; Pilipczuk et al., 2015; Şen & Gökmen, 2022). In contrast, analysis of green beans allows more direct assessment of biochemical changes arising from digestive fermentation, particularly in primary and secondary metabolites (Ifmalinda et al., 2019; Mitra et al., 2025).

Metabolomic profiling using GC-MS offers a suitable analytical approach for this purpose (Frolova et al., 2026) due to its high sensitivity, reproducibility, and effectiveness in detecting low-molecular-weight metabolites, including organic acids, amino acids, sugars, and fatty acid derivatives that are often associated with fermentation pathways (Diez-Simon, Mumm & Hall, 2019; Halket et al., 2005; Pereira et al., 2025). GC-MS also benefits from well-established spectral libraries and robust compound identification, enabling reliable comparison of metabolite profiles between sample groups (Frolova et al., 2026).

In this study, we performed a comparative metabolomic analysis of unfermented coffee beans and green bean civet coffee produced in Vietnam. By characterizing the metabolite profiles of these two types of coffee, we aim to identify metabolic differences associated with the digestive fermentation process and to explore potential metabolite markers that may contribute to the discrimination and authentication of civet coffee.

Material and Methods

Sample collection

Ten green civet coffee and nine normal coffee samples of C. arabica were collected in April 2024 from Me Linh Coffee farm, Da Lat City, Lam Dong Province, Vietnam. Civets were housed individually in separate cages within the farm and maintained under routine daily cleaning. During the coffee harvest season, the animals were fed bananas and rice mixed with meat and vegetables during the daytime and fresh coffee cherries at night. Civet coffee beans were collected the following morning after feeding. Because civet coffee samples were pooled across animals, each sample represented a single-day collection batch; the same sampling approach was applied to normal coffee. As no experimental procedures were performed on animals and the farm is officially registered for coffee production under Me Linh Coffee Limited Company, ethical approval for animal experimentation was not required.

Chemical preparation

Methanol (MeOH, ³ 99.9%), camphor-10-sulfonic acid (b) (CSA, 98%), methoxyamine hydrochloride (98%), pyridine (³ 99.9%), and N-methyl-N-trimethylsilyl trifluoroacetamide (MSTFA, ³ 98.5%) were purchased from Merck (Darmstadt, Germany).

A stock solution of CSA was prepared by dissolving 40 mg of CSA in 50 mL of 80% MeOH to obtain a stock concentration of approximately 3.5 mM. Then, working solutions were prepared by dilution the stock solution with 80% MeOH to obtain the concentration of 210 nM.

Derivatization for GC-MS analysis

Each coffee sample was ground finely and filtered through an 0.22 µm filter. Approximately 50 mg of coffee powder was extracted using 1 mL of 80% MeOH containing 210 nM of CSA as an internal standard and sonicated for 15 min at 37 °C. The solution was incubated at 65 oC in 45 min. Afterward, 300 mL of the extracted sample was evaporated at 70 °C for 1 hr using the SpeedVac Vacuum Concentrator. To maintain the carbonyl moieties, the sample residues were methoxymated using 50 mL of methoxyamine hydrochloride solution (20 mg/mL in pyridine) and incubated for 2 hr at 37 oC with constant agitation at 300 rpm. The samples were then trimethylsilylated by adding 70 μl of MSTFA and incubated with the same settings for another 30 min.

GC-MS analysis

GC/MS analysis was performed using a Trace1310 coupled to an ISQ7000 mass spectrometer (Thermo Scientific, Singapore) equipped with a TriPlus 100 LS autosampler. Separation was achieved using a DB-WAX capillary column (30 m × 0.25 mm × 0.25 µm; Agilent Technologies, California, USA). Samples were injected in split mode (40:1) with the injector temperature maintained at 250 °C. Helium was used as the carrier gas at a constant flow rate of 1.2 mL/min. The oven temperature program was as follows: 80 °C for 2 min, increased to 180 °C at 10 °C/min and held for 2 min, then ramped to 290 °C at 5 °C/min and held for 10 min. The total run time was approximately 46 min. The maximum oven temperature was set to 350 °C with an equilibration time of 0.5 min before each run. Data were acquired in full-scan mode over an m/z range of 85-500.

MS-DIAL processing

Raw GC-MS data (cdf) were converted into abf format using Reifycs Analysis Base File Converter (https://www.reifycs.com/abfconverter/) and processed in MS-DIAL version 5.5.251021 for peak detection, spectrum deconvolution and identification (Takeda et al., 2024). Peaks with intensity greater than 1000 were retained. Signal smoothing was performed using a linear weighted moving average method (level 3; peak width 20 scans), with a sigma window value of 0.6, and an EI spectra threshold of 1. Metabolites annotation was based on retention index and mass spectral matching using tolerances of 15 for retention index, 0.5 min for retention time, and 0.5 Da for m/z. Minimum weighted dot product, dot product, reverse dot product, and identification scores were set to 600, 600, 800 and 700, respectively. Feature alignment was conducted based on retention index matching, and only metabolites detected in at least 50% of samples within each group were retained for subsequent analysis.

Multivariate analysis

Multivariate analysis was performed in R using FactoMineR version 2.12 for principal component analysis (PCA) and ropls version 3.22 for partial least squares-discrimination analysis (PLS-DA) and orthogonal partial least squares-discrimination analysis (OPLS-DA) (Thévenot et al., 2015). Model performance was evaluated by 500 permutation tests with 7-fold cross-validation. Variable importance in projection (VIP) scores were used to estimate variable contribution, with VIP > 1.0 considered significant. Differential metabolite was identified using volcano plot analysis based on false discovery rate (FDR) and log2 fold change (log2FC). Metabolites with FDR < 0.05 and log2FC ³ 1.0 were considered upregulated, whereas those with FDR < 0.05 and log2FC £ -1.0 were considered downregulated. Hierarchical clustering analysis was subsequently performed using Euclidean distance and the single linkage method.

Correlation analysis and pathway enrichment

Pairwise metabolite correlations were calculated using Spearman’s rank correlation. Only correlations with coefficients > 0.70 and Benjamini-Hochberg adjusted p-value < 0.05 were considered significant. Pathway enrichment analysis was performed using the KEGGREST package in R based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) database.

Statistical analysis

Data distribution was assessed using the Kolmogorov-Smirnov test, and between-group differences were evaluated using two-sided Student’s t-test with p-value < 0.05 considered statistically significant. All statistical analyses were implemented in R version 4.4.1 (R Core Team, 2021).

Results and discussion

Untargeted metabolomic analysis of normal and civet coffee bean samples

GC-MS analysis identified a total of 46 metabolites, and detailed peak areas are provided in Table 1. Differential analysis using Student’s t-test combined with fold-change criteria identified 26 metabolites that differed significantly between civet and unprocessed green coffee beans (Table 2). Among these, mannitol was the only metabolite upregulated in civet coffee, whereas all remaining metabolites were downregulated.

Multivariate analysis of metabolite data

To evaluate metabolomic discrimination between the two groups, multivariate analyses were performed using the identified metabolite dataset (Figure 1). Principal component analysis (PCA) showed that PC1 and PC2 explained 53.7% and 21.0% of the total variance, respectively, and clearly separated civet and unprocessed coffee samples without obvious outliers. Civet coffee samples also exhibited greater dispersion, indicating higher variability in metabolite composition. Several metabolites, including ribitol, 2-hydroxybutyric acid, arabinose, glucose, and threonic acid, contributed strongly to sample separation.

The PCA results were further supported by partial least squares discriminant analysis (PLS-DA), in which the first two components explained 53.2% and 17.4% of the variance, respectively (Figure 1C). Variable importance in projection (VIP) analysis identified 24 metabolites with VIP values greater than 1.0, including methylsuccinic acid, glyceric acid, galactose, and glucose (Figure 1D). Orthogonal PLS-DA (OPLS-DA) further improved group discrimination, yielding high explanatory and predictive performance (R2Y = 0.976; Q2 = 0.934) (Figure 1E). Model robustness was confirmed by permutation testing (p < 0.01) (Figure 1F). Combining the statistically significant metabolites with OPLS-DA markers identified 23 overlapping metabolites as potential quality markers, including methylsuccinic acid, galactose, glyceric acid, glucose, arabinose, threonic acid, 3-methylglutaric acid, malic acid, lactic acid, glycine, erythritol, linolenic acid, palmitic acid, citric acid, and mannitol.

Quality markers and pathway analysis

The volcano plot illustrates the distribution of significantly upregulated and downregulated metabolites in civet coffee relative to unprocessed green coffee samples (Figure 2). Hierarchical clustering and box-and-whisker plots of the potential quality markers further revealed both intra- and inter-group variation in metabolite abundance, while clearly separating the two groups (Figure 2B and Figure 3). Among those 23 identified quality markers, only mannitol was upregulated in civet coffee, whereas all other metabolites were downregulated. Notably, several metabolites, including galactose, glucose, and glyceric acid, were nearly absent in civet samples. On the contrary, mannitol was negligible in unprocessed green coffee but remarkably elevated in civet coffee.

Table 1:
Peak area of metabolites in normal and civet samples.

No	Metabolites	Normal samples
		Normal1	Normal2	Normal3	Normal4	Normal5	Normal6	Normal7	Normal8	Normal9
1	Ribitol	15011944	11454823	14246230	8953300	11019178	12578221	7507293	11995909	13564419
2	2-Hydroxybutanoic acid	2488057	2657142	2590400	3039007	3182836	1901379	2371712	2052340	2539897
3	3,6-Anhydro-D-galactose	4883005	3792503	4295477	2663769	3537185	3646706	2423043	3398538	3753640
4	3-Hydroxybenzoate	10272927	11335319	9714898	7309896	9122422	8609483	5880923	14779351	14101886
5	3-Hydroxybutanoic acid	36806253	47724517	40090923	49499089	50230049	29269489	45061831	39003913	43699883
6	Alanine	2454828	3713023	2694295	2786244	2709994	2790570	2011084	3387280	4968634
7	Arabinose	15513913	11578671	15328487	8903081	11576278	14462721	7222042	10816631	12358231
8	beta-Alanine	340224	282014	344278	225156	266490	276102	163381	195091	296009
9	Butane-2,3-diol	5629752	4301333	4466048	4875648	6095959	4406370	4016791	3424650	5230041
10	Citric acid	1.82E+08	2.2E+08	1.58E+08	55335660	1.09E+08	1.55E+08	87435139	1.19E+08	1.76E+08
11	D-Fructose	7446794	5299269	7050004	4553820	5622299	7621947	4362624	6235417	5659070
12	D-Mannitol	11967603	8007628	10412016	5236282	7158783	9801453	6310647	8846911	8152235
13	D-Ribose	62674	46218	54111	33051	45401	55400	24885	44201	53708
14	D-Xylulose	701477	813309	702138	610176	692206	630270	595340	937999	889537
15	Erythritol	1383849	1304597	1525736	1139272	1330117	1358804	790439	1033624	1278601
16	Ethanolamine	17182026	15743260	16324154	13162353	14578258	12873032	9093803	9399921	14527159
17	Fucose	9656968	8790950	10282287	5609674	7989184	10175052	5480961	7803628	8602514
18	Galactose	7.77E+08	5.83E+08	7.26E+08	3.28E+08	4.86E+08	7.29E+08	3.82E+08	6.02E+08	5.69E+08
19	Glucose	7.77E+08	5.83E+08	7.26E+08	3.28E+08	4.86E+08	7.29E+08	3.82E+08	6.02E+08	5.69E+08
20	gamma-Aminobutyric acid (GABA)	11308490	10431104	10913977	6910628	8946097	9986070	5492835	7030159	9825444
21	Glyceric acid	18954896	16114889	17476999	13183725	14380853	13044217	10257722	12521573	15852662
22	Glycine	9182045	10192260	9106217	7602656	9009017	9271928	5869435	6210841	9756708
23	Iminodiacetate	41956442	25589793	32232488	12799898	20699295	25662406	16700665	16751355	20499305
24	Inositol	3.49E+08	2.97E+08	3.36E+08	1.49E+08	2.44E+08	3.69E+08	1.98E+08	2.84E+08	2.81E+08
25	Isocitric acid	9106218	7024807	7681058	4396553	7175534	6277018	4189016	8283422	8223985
26	Lactic acid	7516780	9131593	9704271	9713339	9087487	8471731	5247854	5654137	7244638
27	Linolenic acid	66331133	42894711	60290924	19652593	33560579	53105956	20872690	45385406	38404881
28	L-Norleucine	8169138	5083450	7667545	4176460	4741582	6109895	3561924	3290144	4553146
29	L-Phenylalanine	8739373	13726246	9748335	10316405	7016921	6432543	5671494	9919311	14969505
30	Malic acid	1E+09	1.15E+09	9.53E+08	6.11E+08	8.87E+08	9.51E+08	5.64E+08	7.15E+08	1.1E+09
31	N-Acetyl-D-glucosamine	3.12E+08	2.49E+08	3.06E+08	1.38E+08	2.16E+08	3.14E+08	1.7E+08	2.58E+08	2.43E+08
32	n-Heptadecane	658536	605266	677126	662265	617917	530890	525910	502394	613771
33	Nicotinic acid	17283985	12190162	18684880	11413785	20629792	13872928	8544137	13523856	21458449
34	Oxalic acid	797724	513460	745212	834384	665717	499036	539348	450582	533731
35	Palmitic acid	1.09E+08	83569496	1.01E+08	43193339	67453515	91121827	47520954	93609630	81514649
36	Pipecolic acid	3715269	5686249	3391291	1690782	4237905	2530179	2535987	4331487	4409396
37	Proline	37036323	10208667	28331637	11086261	12327362	17487855	10135165	8723713	11204801
38	Pyroglutamic acid	12242985	13968933	11289222	6557581	8340021	11270676	6310996	7895965	13348770
39	Quinic acid	7.27E+08	6.04E+08	7E+08	2.56E+08	5.13E+08	7.85E+08	3.5E+08	6.39E+08	6.03E+08
40	Sulfuric acid	1299475	1175760	1378416	1537897	1655990	1101184	1296500	1047068	1224916
41	Threitol	2339036	1872474	2127065	1467068	1939141	1919898	1222022	1301505	1746848
42	Threonic acid	36907089	27941304	31269153	19327063	25698429	24659896	16011858	24144030	27890097
43	Threonine	8111166	4713742	7347397	3676566	4329623	4913096	3788891	4620787	4357757
44	Trigonelline	17412998	17791613	15276052	12512619	14653791	15645812	4283664	4310484	17603027
45	Methyl succinic acid	25094416	23462417	25980272	21784456	24154981	22472557	14134484	15582053	23719369
46	2-deoxytetronic acid	9944933	9888315	9376362	7645502	8687050	8491498	5615465	5260582	10086246
No	Metabolites	Civet samples
		Civet1	Civet2	Civet3	Civet4	Civet5	Civet6	Civet7
1	Ribitol	4062652	2696230	3454617	23407324	5642068	4198638	6075514
2	2-Hydroxybutanoic acid	3611822	676795	393974	12557	894782	414911	681852
3	3,6-Anhydro-D-galactose	2443229	1000181	1098427	3431285	2331571	1713016	2233562
4	3-Hydroxybenzoate	19161544	8576708	6788550	39235410	26066769	24900840	17284862
5	3-Hydroxybutanoic acid	44938035	37040213	39116912	1853157	52100873	53466697	53331207
6	Alanine	2946749	1072529	1310899	3307572	2857422	2743838	4011891
7	Arabinose	1214238	802205	826110	2924000	1231443	1037499	743171
8	beta-Alanine	331657	183051	195889	16393	329942	211551	312679
9	Butane-2,3-diol	6889311	3819142	3453920	5817670	6388233	6122311	5199255
10	Citric acid	1E+08	54472965	72741333	1799461	1.02E+08	62438714	1.29E+08
11	D-Fructose	6695169	4437634	4071082	12335842	7342362	4835917	6982731
12	D-Mannitol	79140218	42340588	51244485	86314557	70377486	42541750	73223669
13	D-Ribose	39364	14320	22318	47986	37074	18044	27756
14	D-Xylulose	281148	365512	391172	784267	1002455	1083572	798633
15	Erythritol	868079	470896	516346	804535	770125	531022	715981
16	Ethanolamine	10354022	7248377	8482451	5121	11170332	8623742	11162064
17	Fucose	10944399	7697144	9281632	6812283	12914080	9247872	12324700
18	Galactose	12526644	7618246	7844023	21180063	13354489	9608745	13426909
19	Glucose	26178022	15574091	17342362	44922248	27484290	20873620	28003534
20	gamma-Aminobutyric acid (GABA)	31281932	19344874	24773533	12109	35475629	21733279	36683869
21	Glyceric acid	3376532	2717896	2354338	3523736	4734690	3801855	4200434
22	Glycine	4680478	2651639	3086320	7918	4776348	3147277	4387139
23	Iminodiacetate	35870422	3143899	5318326	4286390	6807278	3804806	6803170
24	Inositol	4.23E+08	2.16E+08	2.72E+08	6.05E+08	3.76E+08	2.3E+08	4.13E+08
25	Isocitric acid	10052993	5172027	4923782	10383114	11238633	8664312	9572378
26	Lactic acid	4706759	2375762	2587059	4198401	3533013	2628120	3609107
27	Linolenic acid	14751854	7045664	9184646	27098318	18529693	10018093	15871938
28	L-Norleucine	297370	1064155	2088934	5986	1824460	1158809	1980603
29	L-Phenylalanine	2562861	2195152	2352117	6729720	4007666	2647662	3700682
30	Malic acid	3.79E+08	1.97E+08	2.4E+08	15141605	3.75E+08	2.24E+08	3.98E+08
31	N-Acetyl-D-glucosamine	1.42E+08	71830775	88551132	1.85E+08	1.24E+08	80868259	1.3E+08
32	n-Heptadecane	450741	462705	448922	788375	962041	618605	583233
33	Nicotinic acid	32587174	5050040	15410947	52323	45846382	40982030	40593822
34	Oxalic acid	1139267	748196	732668	733467	1026338	922874	922217
35	Palmitic acid	58111231	26487842	27101373	59361893	54469663	30829479	41830873
36	Pipecolic acid	8456799	3003879	2614092	3811166	6411587	3701056	5412452
37	Proline	117539	1922887	7184662	22931	4196701	3942323	4901892
38	Pyroglutamic acid	6847344	4072066	5745487	437213	11947261	6499855	9528803
39	Quinic acid	6.54E+08	2.69E+08	3.5E+08	1.02E+08	5.24E+08	2.44E+08	6E+08
40	Sulfuric acid	992767	659381	768347	503782	987561	769789	805820
41	Threitol	1628185	854437	1057032	1624602	1508334	1175317	1538111
42	Threonic acid	3594252	2064629	1892011	4987081	4239406	3392634	3815529
43	Threonine	1546836	1729354	2566674	3577607	3459372	2072333	2521497
44	Trigonelline	11076653	7381828	8517842	2197389	12026675	5259199	16022046
45	Methyl succinic acid	6254708	2911777	3389298	1234604	5675863	3924230	4886743
46	2-deoxytetronic acid	4104733	2065501	2797409	618972	4568217	3604655	3814728

Table 2:
Statistical analysis and comparative evaluation of identified metabolites between unprocessed and civet coffee green beans.

No	Name	p-value	FDR*	log2FC**	Status
1	Galactose	2.76E-15	1.27E-13	-6.09	Down regulated
2	Glucose	1.13E-14	2.61E-13	-5.87	Down regulated
3	Glyceric acid	4.10E-12	6.29E-11	-2.80	Down regulated
4	Threonic acid	1.38E-07	1.59E-06	-2.96	Down regulated
5	Threonine	4.48E-06	3.94E-05	-1.53	Down regulated
6	Linolenic acid	5.13E-06	3.94E-05	-1.99	Down regulated
7	Arabinose	2.90E-05	1.91E-04	-2.29	Down regulated
8	Urea	1.02E-04	5.84E-04	-1.24	Down regulated
9	Palmitic acid	1.41E-04	7.19E-04	-1.46	Down regulated
10	Methylsuccinic acid	1.83E-04	8.40E-04	-3.27	Down regulated
11	Aspartic acid	3.86E-04	1.61E-03	-1.55	Down regulated
12	Mannitol	4.22E-04	1.62E-03	2.97	Up regulated
13	3-Methylglutaric acid	5.03E-04	1.78E-03	-2.09	Down regulated
14	Lactic acid	7.62E-04	2.51E-03	-1.68	Down regulated
15	N-Acetyl-D-glucosamine	9.99E-04	3.06E-03	-1.14	Down regulated
16	Ribitol	1.18E-03	3.39E-03	-1.24	Down regulated
17	Proline	1.45E-03	3.93E-03	-2.91	Down regulated
18	L-Norleucine	2.24E-03	5.72E-03	-2.70	Down regulated
19	2-Hydroxybutyric acid	3.09E-03	6.98E-03	-1.95	Down regulated
20	Malic acid	3.17E-03	6.98E-03	-2.33	Down regulated
21	Threitol	3.19E-03	6.98E-03	-0.73	Not significant
22	Erythritol	6.26E-03	1.31E-02	-1.40	Down regulated
23	D-Ribose	8.26E-03	1.65E-02	-0.92	Not significant
24	Glycine	8.97E-03	1.72E-02	-1.94	Down regulated
25	Citric acid	1.39E-02	2.57E-02	-1.47	Down regulated
26	Pyroglutamic acid	1.76E-02	3.12E-02	-1.16	Down regulated
27	Trigonelline	2.22E-02	3.76E-02	-1.03	Down regulated
28	Ethanolamine	2.29E-02	3.76E-02	-1.27	Down regulated
29	3-Hydroxybenzoic acid	2.47E-02	3.92E-02	0.84	Not significant
30	beta-Alanine	3.79E-02	5.81E-02	-0.67	Not significant
31	Inositol	4.08E-02	6.06E-02	0.58	Not significant
32	3-Hydroxybutanoic acid	4.53E-02	6.52E-02	-0.63	Not significant
33	Quinic acid	5.05E-02	7.04E-02	-1.06	Not significant
34	L-Phenylalanine	5.65E-02	7.65E-02	-0.40	Not significant
35	n-Heptadecane	6.76E-02	8.89E-02	0.50	Not significant
36	Alanine	9.43E-02	1.20E-01	-0.29	Not significant
37	Nicotinic acid	1.09E-01	1.35E-01	0.07	Not significant
38	Pipecolic acid	1.24E-01	1.51E-01	0.51	Not significant
39	gamma-Aminobutyric acid (GABA)	1.93E-01	2.27E-01	0.97	Not significant
40	Isocitric acid	2.00E-01	2.30E-01	0.27	Not significant
41	3,6-Anhydro-D-galactose	2.50E-01	2.80E-01	-0.12	Not significant
42	D-Fructose	2.65E-01	2.91E-01	-0.16	Not significant
43	Butane-2,3-diol	2.75E-01	2.94E-01	0.19	Not significant
44	D-Xylulose	2.85E-01	2.98E-01	-0.16	Not significant
45	Fucose	5.54E-01	5.66E-01	-0.05	Not significant
46	Oxalic acid	9.32E-01	9.32E-01	0.17	Not significant

• Not significant: FDR ≥ 0.05 and/or |log2FC| <1, Down regulated: FDR <0.05 and log2FC ≤ -1, Up regulated: FDR < 0.05 and log2FC ≥ 1, *: False discovery rate; **: Fold change

Figure 1:
Multivariate analysis of civet and unprocessed green coffee samples. (A) PCA plot of the first two principal components; (B) PCA variable correlation plot, arrow colors are scaled according to the cos2 values of the corresponding metabolites; (C) PLS-DA score plot; (D) Variable importance in projection (VIP) scores of discriminant metabolites; (E) OPLS-DA score plot; (F) Permutation test results for OPLS-DA model validation; (G) S-plot derived from the OPLS-DA model.

Figure 2:
Volcano plot and hierarchical clustering of metabolites in civet and unprocessed green coffee samples. (A) Volcano plot of all identified metabolites; (B) Heatmap of differential metabolites. Metabolites were clustered using hierarchical clustering analysis and annotated according to their subclass in the KEGG databases.

Figure 3:
Box-and-whisker plots of potential quality markers. The red boxes represent the civet coffee group, while the blue boxed represents the unprocessed green coffee group. The median peak intensity is indicated by the central line within each box, and the 5^th and 95^th percentiles are shown at the lower and upper bounds, respectively. Individual data points corresponding to each sample are overlaid on the plots. Statistical comparisons between the two groups were performed using a two-sided Student’s t-test, and significance levels are indicated above each plot (*: p<0.05, **: p<0.01, ***: p<0.001, ****: p<0.0001).

Figure 4:
Correlation network and KEGG enrichment analysis. (A) Correlation network of all identified metabolites. Nodes represent metabolites clustered into four modules by Louvain algorithm and are colored accordingly. Edges indicate the direction of correlation between metabolite pair: positive (red) and negative (blue). Potential quality markers identified from OPLS-DA are highlighted with triangular shape. (B) Pathway enrichment analysis of potential quality markers based on the KEGG database. Point color corresponds to the adjusted p-value, while point size reflects the number of metabolites enriched in each pathway. (C) Two KEGG pathways with the most extensively characterized metabolic reactions in C. arabica. Green arrows indicate reactions that have been reported in C. arabica. Red circular nodes represent enriched metabolites, whereas white nodes and black arrows denote compounds and reactions from the global KEGG metabolic maps.

To further elucidate the relationships among identified metabolites, a metabolite correlation network was constructed, followed by pathway enrichment analysis of the 23 quality markers using the KEGG database (Figure 4; Table 3). Most metabolites exhibited strong positive correlations, whereas only a few pairs showed negative correlations, notably oxalic acid with 3,6-anhydro-D-galactose, and threonic acid with mannitol. In addition, network analysis revealed that the largest cluster (module 1) contained a substantial proportion of quality markers, while the remaining markers were primarily grouped within a second major cluster (module 2). These modules were strongly interconnected (Figure 4), which suggests a coordinated and global metabolic shift in coffee beans associated with exposure to the civet gastrointestinal tract.

Table 3:
Correlation analysis results.

Number	From	To	r	Number	From	To	r
1	D-AnG	Ribitol	0.77	395	2-HB	L-Norleucine	0.69
2	Alanine	Ribitol	0.71	396	Arabinose	L-Norleucine	0.74
3	Arabinose	Ribitol	0.95	397	beta-Alanine	L-Norleucine	0.67
4	D-Fructose	Ribitol	0.75	398	Citric acid	L-Norleucine	0.86
5	D-Ribose	Ribitol	0.68	399	D-Ribose	L-Norleucine	0.72
6	Erythritol	Ribitol	0.70	400	Erythritol	L-Norleucine	0.85
7	Galactose	Ribitol	0.81	401	Ethanolamine	L-Norleucine	0.91
8	Glucose	Ribitol	0.78	402	Galactose	L-Norleucine	0.83
9	Glyceric acid	Ribitol	0.86	403	Glucose	L-Norleucine	0.85
10	Aspartic acid	Ribitol	0.69	404	Glyceric acid	L-Norleucine	0.84
11	Linolenic acid	Ribitol	0.81	405	Glycine	L-Norleucine	0.92
12	L-Phenylalanine	Ribitol	0.78	406	Aspartic acid	L-Norleucine	0.74
13	NAG	Ribitol	0.85	407	Lactic acid	L-Norleucine	0.84
14	Palmitic acid	Ribitol	0.68	408	Linolenic acid	L-Norleucine	0.85
15	Threonic acid	Ribitol	0.93	409	Malic acid	L-Norleucine	0.93
16	Threonine	Ribitol	0.89	410	NAG	L-Norleucine	0.69
17	Methylsuccinic acid	Ribitol	0.66	411	Palmitic acid	L-Norleucine	0.85
18	beta-Alanine	2-HB	0.66	412	Proline	L-Norleucine	0.99
19	Citric acid	2-HB	0.66	413	Pyroglutamic acid	L-Norleucine	0.81
20	Erythritol	2-HB	0.79	414	Quinic acid	L-Norleucine	0.73
21	Ethanolamine	2-HB	0.82	415	Urea	L-Norleucine	0.88
22	Glycine	2-HB	0.80	416	Threitol	L-Norleucine	0.78
23	Aspartic acid	2-HB	0.77	417	Threonic acid	L-Norleucine	0.76
24	Lactic acid	2-HB	0.83	418	Threonine	L-Norleucine	0.82
25	L-Norleucine	2-HB	0.69	419	Trigonelline	L-Norleucine	0.80
26	Malic acid	2-HB	0.79	420	Methylsuccinic acid	L-Norleucine	0.94
27	Palmitic acid	2-HB	0.73	421	3MG	L-Norleucine	0.93
28	Proline	2-HB	0.69	422	Ribitol	L-Phenylalanine	0.78
29	Pyroglutamic acid	2-HB	0.70	423	D-AnG	L-Phenylalanine	0.98
30	Urea	2-HB	0.89	424	Alanine	L-Phenylalanine	0.80
31	Threitol	2-HB	0.71	425	Arabinose	L-Phenylalanine	0.71
32	Trigonelline	2-HB	0.67	426	D-Fructose	L-Phenylalanine	0.69
33	Methylsuccinic acid	2-HB	0.84	427	Threonic acid	L-Phenylalanine	0.73
34	3MG	2-HB	0.82	428	Threonine	L-Phenylalanine	0.66
35	Ribitol	D-AnG	0.77	429	2-HB	Malic acid	0.79
36	Alanine	D-AnG	0.80	430	Arabinose	Malic acid	0.72
37	Arabinose	D-AnG	0.69	431	beta-Alanine	Malic acid	0.70
38	D-Fructose	D-AnG	0.77	432	Citric acid	Malic acid	0.92
39	L-Phenylalanine	D-AnG	0.98	433	D-Ribose	Malic acid	0.77
40	Oxalic acid	D-AnG	-0.67	434	Erythritol	Malic acid	0.89
41	Threonic acid	D-AnG	0.68	435	Ethanolamine	Malic acid	0.93
42	Inositol	3-HBA	0.68	436	Galactose	Malic acid	0.87
43	Isocitric acid	3-HBA	0.84	437	Glucose	Malic acid	0.89
44	n-Heptadecane	3-HBA	0.68	438	Glyceric acid	Malic acid	0.85
45	Pipecolic acid	3-HBA	0.75	439	Glycine	Malic acid	0.97
46	Ribitol	Alanine	0.71	440	Aspartic acid	Malic acid	0.79
47	D-AnG	Alanine	0.80	441	Lactic acid	Malic acid	0.88
48	L-Phenylalanine	Alanine	0.80	442	Linolenic acid	Malic acid	0.87
49	Ribitol	Arabinose	0.95	443	L-Norleucine	Malic acid	0.93
50	D-AnG	Arabinose	0.69	444	NAG	Malic acid	0.75
51	D-Ribose	Arabinose	0.66	445	Palmitic acid	Malic acid	0.91
52	Erythritol	Arabinose	0.75	446	Proline	Malic acid	0.90
53	Ethanolamine	Arabinose	0.66	447	Pyroglutamic acid	Malic acid	0.90
54	Galactose	Arabinose	0.84	448	Quinic acid	Malic acid	0.78
55	Glucose	Arabinose	0.83	449	Urea	Malic acid	0.89
56	Glyceric acid	Arabinose	0.93	450	Threitol	Malic acid	0.85
57	Glycine	Arabinose	0.73	451	Threonic acid	Malic acid	0.76
58	Aspartic acid	Arabinose	0.69	452	Threonine	Malic acid	0.76
59	Lactic acid	Arabinose	0.70	453	Trigonelline	Malic acid	0.84
60	Linolenic acid	Arabinose	0.86	454	Methylsuccinic acid	Malic acid	0.96
61	L-Norleucine	Arabinose	0.74	455	3MG	Malic acid	0.98
62	L-Phenylalanine	Arabinose	0.71	456	Ribitol	NAG	0.85
63	Malic acid	Arabinose	0.72	457	Arabinose	NAG	0.81
64	NAG	Arabinose	0.81	458	Citric acid	NAG	0.67
65	Palmitic acid	Arabinose	0.72	459	D-Ribose	NAG	0.88
66	Proline	Arabinose	0.74	460	Erythritol	NAG	0.84
67	Urea	Arabinose	0.68	461	Galactose	NAG	0.94
68	Threitol	Arabinose	0.66	462	Glucose	NAG	0.92
69	Threonic acid	Arabinose	0.98	463	Glyceric acid	NAG	0.74
70	Threonine	Arabinose	0.95	464	Glycine	NAG	0.69
71	Methylsuccinic acid	Arabinose	0.75	465	Aspartic acid	NAG	0.67
72	3MG	Arabinose	0.74	466	Lactic acid	NAG	0.75
73	2-HB	beta-Alanine	0.66	467	Linolenic acid	NAG	0.92
74	Citric acid	beta-Alanine	0.82	468	L-Norleucine	NAG	0.69
75	Ethanolamine	beta-Alanine	0.85	469	Malic acid	NAG	0.75
76	Fucose	beta-Alanine	0.86	470	Palmitic acid	NAG	0.86
77	GABA	beta-Alanine	0.72	471	Proline	NAG	0.67
78	Glycine	beta-Alanine	0.69	472	Quinic acid	NAG	0.70
79	L-Norleucine	beta-Alanine	0.67	473	Threitol	NAG	0.78
80	Malic acid	beta-Alanine	0.70	474	Threonic acid	NAG	0.78
81	Nicotinic acid	beta-Alanine	0.76	475	Threonine	NAG	0.76
82	Palmitic acid	beta-Alanine	0.72	476	Methylsuccinic acid	NAG	0.74
83	Proline	beta-Alanine	0.67	477	3MG	NAG	0.73
84	Pyroglutamic acid	beta-Alanine	0.87	478	3-HBA	n-Heptadecane	0.68
85	Quinic acid	beta-Alanine	0.80	479	beta-Alanine	Nicotinic acid	0.76
86	Threitol	beta-Alanine	0.70	480	Fucose	Nicotinic acid	0.84
87	Trigonelline	beta-Alanine	0.89	481	GABA	Nicotinic acid	0.89
88	Methylsuccinic acid	beta-Alanine	0.74	482	Oxalic acid	Nicotinic acid	0.80
89	3MG	beta-Alanine	0.73	483	D-AnG	Oxalic acid	-0.67
90	Oxalic acid	Butane-2,3-diol	0.69	484	Butane-2,3-diol	Oxalic acid	0.69
91	2-HB	Citric acid	0.66	485	Fucose	Oxalic acid	0.67
92	beta-Alanine	Citric acid	0.82	486	GABA	Oxalic acid	0.79
93	D-Ribose	Citric acid	0.69	487	Nicotinic acid	Oxalic acid	0.80
94	Erythritol	Citric acid	0.75	488	Ribitol	Palmitic acid	0.68
95	Ethanolamine	Citric acid	0.89	489	2-HB	Palmitic acid	0.73
96	Fucose	Citric acid	0.67	490	Arabinose	Palmitic acid	0.72
97	Galactose	Citric acid	0.77	491	beta-Alanine	Palmitic acid	0.72
98	Glucose	Citric acid	0.79	492	Citric acid	Palmitic acid	0.87
99	Glyceric acid	Citric acid	0.76	493	D-Ribose	Palmitic acid	0.88
100	Glycine	Citric acid	0.90	494	Erythritol	Palmitic acid	0.92
101	Aspartic acid	Citric acid	0.77	495	Ethanolamine	Palmitic acid	0.85
102	Lactic acid	Citric acid	0.68	496	Galactose	Palmitic acid	0.92
103	Linolenic acid	Citric acid	0.78	497	Glucose	Palmitic acid	0.94
104	L-Norleucine	Citric acid	0.86	498	Glyceric acid	Palmitic acid	0.79
105	Malic acid	Citric acid	0.92	499	Glycine	Palmitic acid	0.87
106	NAG	Citric acid	0.67	500	Aspartic acid	Palmitic acid	0.81
107	Palmitic acid	Citric acid	0.87	501	Lactic acid	Palmitic acid	0.83
108	Proline	Citric acid	0.82	502	Linolenic acid	Palmitic acid	0.92
109	Pyroglutamic acid	Citric acid	0.94	503	L-Norleucine	Palmitic acid	0.85
110	Quinic acid	Citric acid	0.86	504	Malic acid	Palmitic acid	0.91
111	Urea	Citric acid	0.73	505	NAG	Palmitic acid	0.86
112	Threitol	Citric acid	0.76	506	Proline	Palmitic acid	0.83
113	Threonic acid	Citric acid	0.67	507	Pyroglutamic acid	Palmitic acid	0.84
114	Threonine	Citric acid	0.69	508	Quinic acid	Palmitic acid	0.89
115	Trigonelline	Citric acid	0.88	509	Urea	Palmitic acid	0.81
116	Methylsuccinic acid	Citric acid	0.86	510	Threitol	Palmitic acid	0.88
117	3MG	Citric acid	0.89	511	Threonic acid	Palmitic acid	0.72
118	Ribitol	D-Fructose	0.75	512	Threonine	Palmitic acid	0.73
119	D-AnG	D-Fructose	0.77	513	Trigonelline	Palmitic acid	0.75
120	L-Phenylalanine	D-Fructose	0.69	514	Methylsuccinic acid	Palmitic acid	0.93
121	Threonic acid	Mannitol	-0.67	515	3MG	Palmitic acid	0.92
122	Ribitol	D-Ribose	0.68	516	3-HBA	Pipecolic acid	0.75
123	Arabinose	D-Ribose	0.66	517	Inositol	Pipecolic acid	0.66
124	Citric acid	D-Ribose	0.69	518	Isocitric acid	Pipecolic acid	0.80
125	Erythritol	D-Ribose	0.92	519	2-HB	Proline	0.69
126	Ethanolamine	D-Ribose	0.74	520	Arabinose	Proline	0.74
127	Galactose	D-Ribose	0.87	521	beta-Alanine	Proline	0.67
128	Glucose	D-Ribose	0.86	522	Citric acid	Proline	0.82
129	Glyceric acid	D-Ribose	0.65	523	D-Ribose	Proline	0.69
130	Glycine	D-Ribose	0.70	524	Erythritol	Proline	0.84
131	Lactic acid	D-Ribose	0.80	525	Ethanolamine	Proline	0.90
132	Linolenic acid	D-Ribose	0.85	526	Galactose	Proline	0.81
133	L-Norleucine	D-Ribose	0.72	527	Glucose	Proline	0.82
134	Malic acid	D-Ribose	0.77	528	Glyceric acid	Proline	0.83
135	NAG	D-Ribose	0.88	529	Glycine	Proline	0.90
136	Palmitic acid	D-Ribose	0.88	530	Aspartic acid	Proline	0.73
137	Proline	D-Ribose	0.69	531	Lactic acid	Proline	0.82
138	Pyroglutamic acid	D-Ribose	0.71	532	Linolenic acid	Proline	0.82
139	Quinic acid	D-Ribose	0.76	533	L-Norleucine	Proline	0.99
140	Urea	D-Ribose	0.69	534	Malic acid	Proline	0.90
141	Threitol	D-Ribose	0.93	535	N-Acetyl-D-glucosamine	Proline	0.67
142	Trigonelline	D-Ribose	0.68	536	Palmitic acid	Proline	0.83
143	Methylsuccinic acid	D-Ribose	0.80	537	Pyroglutamic acid	Proline	0.78
144	3MG	D-Ribose	0.78	538	Quinic acid	Proline	0.72
145	Ribitol	Erythritol	0.70	539	Urea	Proline	0.89
146	2-HB	Erythritol	0.79	540	Threitol	Proline	0.74
147	Arabinose	Erythritol	0.75	541	Threonic acid	Proline	0.75
148	Citric acid	Erythritol	0.75	542	Threonine	Proline	0.82
149	D-Ribose	Erythritol	0.92	543	Trigonelline	Proline	0.78
150	Ethanolamine	Erythritol	0.86	544	Methylsuccinic acid	Proline	0.93
151	Galactose	Erythritol	0.90	545	3MG	Proline	0.92
152	Glucose	Erythritol	0.91	546	2-HB	Pyroglutamic acid	0.70
153	Glyceric acid	Erythritol	0.78	547	beta-Alanine	Pyroglutamic acid	0.87
154	Glycine	Erythritol	0.85	548	Citric acid	Pyroglutamic acid	0.94
155	Aspartic acid	Erythritol	0.74	549	D-Ribose	Pyroglutamic acid	0.71
156	Lactic acid	Erythritol	0.94	550	Erythritol	Pyroglutamic acid	0.76
157	Linolenic acid	Erythritol	0.88	551	Ethanolamine	Pyroglutamic acid	0.91
158	L-Norleucine	Erythritol	0.85	552	Fucose	Pyroglutamic acid	0.71
159	Malic acid	Erythritol	0.89	553	Galactose	Pyroglutamic acid	0.72
160	NAG	Erythritol	0.84	554	Glucose	Pyroglutamic acid	0.74
161	Palmitic acid	Erythritol	0.92	555	Glyceric acid	Pyroglutamic acid	0.74
162	Proline	Erythritol	0.84	556	Glycine	Pyroglutamic acid	0.91
163	Pyroglutamic acid	Erythritol	0.76	557	Aspartic acid	Pyroglutamic acid	0.71
164	Quinic acid	Erythritol	0.75	558	Lactic acid	Pyroglutamic acid	0.72
165	Urea	Erythritol	0.89	559	Linolenic acid	Pyroglutamic acid	0.75
166	Threitol	Erythritol	0.94	560	L-Norleucine	Pyroglutamic acid	0.81
167	Threonic acid	Erythritol	0.72	561	Malic acid	Pyroglutamic acid	0.90
168	Threonine	Erythritol	0.71	562	Palmitic acid	Pyroglutamic acid	0.84
169	Trigonelline	Erythritol	0.75	563	Proline	Pyroglutamic acid	0.78
170	Methylsuccinic acid	Erythritol	0.94	564	Quinic acid	Pyroglutamic acid	0.80
171	3MG	Erythritol	0.90	565	Urea	Pyroglutamic acid	0.75
172	2-HB	Ethanolamine	0.82	566	Threitol	Pyroglutamic acid	0.82
173	Arabinose	Ethanolamine	0.66	567	Trigonelline	Pyroglutamic acid	0.90
174	beta-Alanine	Ethanolamine	0.85	568	Methylsuccinic acid	Pyroglutamic acid	0.85
175	Citric acid	Ethanolamine	0.89	569	3MG	Pyroglutamic acid	0.89
176	D-Ribose	Ethanolamine	0.74	570	beta-Alanine	Quinic acid	0.80
177	Erythritol	Ethanolamine	0.86	571	Citric acid	Quinic acid	0.86
178	Galactose	Ethanolamine	0.77	572	D-Ribose	Quinic acid	0.76
179	Glucose	Ethanolamine	0.79	573	Erythritol	Quinic acid	0.75
180	Glyceric acid	Ethanolamine	0.84	574	Ethanolamine	Quinic acid	0.76
181	Glycine	Ethanolamine	0.92	575	Fucose	Quinic acid	0.76
182	Aspartic acid	Ethanolamine	0.79	576	Galactose	Quinic acid	0.74
183	Lactic acid	Ethanolamine	0.85	577	Glucose	Quinic acid	0.76
184	Linolenic acid	Ethanolamine	0.79	578	Glycine	Quinic acid	0.73
185	L-Norleucine	Ethanolamine	0.91	579	Aspartic acid	Quinic acid	0.76
186	Malic acid	Ethanolamine	0.93	580	Linolenic acid	Quinic acid	0.72
187	Palmitic acid	Ethanolamine	0.85	581	L-Norleucine	Quinic acid	0.73
188	Proline	Ethanolamine	0.90	582	Malic acid	Quinic acid	0.78
189	Pyroglutamic acid	Ethanolamine	0.91	583	NAG	Quinic acid	0.70
190	Quinic acid	Ethanolamine	0.76	584	Palmitic acid	Quinic acid	0.89
191	Urea	Ethanolamine	0.91	585	Proline	Quinic acid	0.72
192	Threitol	Ethanolamine	0.85	586	Pyroglutamic acid	Quinic acid	0.80
193	Threonic acid	Ethanolamine	0.72	587	Threitol	Quinic acid	0.74
194	Threonine	Ethanolamine	0.73	588	Trigonelline	Quinic acid	0.79
195	Trigonelline	Ethanolamine	0.90	589	Methylsuccinic acid	Quinic acid	0.77
196	Methylsuccinic acid	Ethanolamine	0.95	590	3MG	Quinic acid	0.75
197	3MG	Ethanolamine	0.95	591	2-HB	Urea	0.89
198	beta-Alanine	Fucose	0.86	592	Arabinose	Urea	0.68
199	Citric acid	Fucose	0.67	593	Citric acid	Urea	0.73
200	GABA	Fucose	0.89	594	D-Ribose	Urea	0.69
201	Nicotinic acid	Fucose	0.84	595	Erythritol	Urea	0.89
202	Oxalic acid	Fucose	0.67	596	Ethanolamine	Urea	0.91
203	Pyroglutamic acid	Fucose	0.71	597	Galactose	Urea	0.78
204	Quinic acid	Fucose	0.76	598	Glucose	Urea	0.80
205	Trigonelline	Fucose	0.71	599	Glyceric acid	Urea	0.80
206	Ribitol	Galactose	0.81	600	Glycine	Urea	0.88
207	Arabinose	Galactose	0.84	601	Aspartic acid	Urea	0.73
208	Citric acid	Galactose	0.77	602	Lactic acid	Urea	0.93
209	D-Ribose	Galactose	0.87	603	Linolenic acid	Urea	0.78
210	Erythritol	Galactose	0.90	604	L-Norleucine	Urea	0.88
211	Ethanolamine	Galactose	0.77	605	Malic acid	Urea	0.89
212	Glucose	Galactose	0.99	606	Palmitic acid	Urea	0.81
213	Glyceric acid	Galactose	0.85	607	Proline	Urea	0.89
214	Glycine	Galactose	0.81	608	Pyroglutamic acid	Urea	0.75
215	Aspartic acid	Galactose	0.74	609	Threitol	Urea	0.79
216	Lactic acid	Galactose	0.85	610	Threonic acid	Urea	0.69
217	Linolenic acid	Galactose	0.99	611	Threonine	Urea	0.68
218	L-Norleucine	Galactose	0.83	612	Trigonelline	Urea	0.75
219	Malic acid	Galactose	0.87	613	Methylsuccinic acid	Urea	0.95
220	NAG	Galactose	0.94	614	3MG	Urea	0.92
221	Palmitic acid	Galactose	0.92	615	2-HB	Threitol	0.71
222	Proline	Galactose	0.81	616	Arabinose	Threitol	0.66
223	Pyroglutamic acid	Galactose	0.72	617	beta-Alanine	Threitol	0.70
224	Quinic acid	Galactose	0.74	618	Citric acid	Threitol	0.76
225	Urea	Galactose	0.78	619	D-Ribose	Threitol	0.93
226	Threitol	Galactose	0.84	620	Erythritol	Threitol	0.94
227	Threonic acid	Galactose	0.84	621	Ethanolamine	Threitol	0.85
228	Threonine	Galactose	0.84	622	Galactose	Threitol	0.84
229	Trigonelline	Galactose	0.66	623	Glucose	Threitol	0.86
230	Methylsuccinic acid	Galactose	0.86	624	Glyceric acid	Threitol	0.74
231	3MG	Galactose	0.85	625	Glycine	Threitol	0.78
232	Ribitol	Glucose	0.78	626	Aspartic acid	Threitol	0.71
233	Arabinose	Glucose	0.83	627	Lactic acid	Threitol	0.86
234	Citric acid	Glucose	0.79	628	Linolenic acid	Threitol	0.84
235	D-Ribose	Glucose	0.86	629	L-Norleucine	Threitol	0.78
236	Erythritol	Glucose	0.91	630	Malic acid	Threitol	0.85
237	Ethanolamine	Glucose	0.79	631	NAG	Threitol	0.78
238	Galactose	Glucose	0.99	632	Palmitic acid	Threitol	0.88
239	Glyceric acid	Glucose	0.85	633	Proline	Threitol	0.74
240	Glycine	Glucose	0.82	634	Pyroglutamic acid	Threitol	0.82
241	Aspartic acid	Glucose	0.76	635	Quinic acid	Threitol	0.74
242	Lactic acid	Glucose	0.86	636	Urea	Threitol	0.79
243	Linolenic acid	Glucose	0.98	637	Threonic acid	Threitol	0.66
244	L-Norleucine	Glucose	0.85	638	Trigonelline	Threitol	0.80
245	Malic acid	Glucose	0.89	639	Methylsuccinic acid	Threitol	0.88
246	NAG	Glucose	0.92	640	3MG	Threitol	0.86
247	Palmitic acid	Glucose	0.94	641	Ribitol	Threonic acid	0.93
248	Proline	Glucose	0.82	642	D-AnG	Threonic acid	0.68
249	Pyroglutamic acid	Glucose	0.74	643	Arabinose	Threonic acid	0.98
250	Quinic acid	Glucose	0.76	644	Citric acid	Threonic acid	0.67
251	Urea	Glucose	0.80	645	Mannitol	Threonic acid	-0.67
252	Threitol	Glucose	0.86	646	Erythritol	Threonic acid	0.72
253	Threonic acid	Glucose	0.82	647	Ethanolamine	Threonic acid	0.72
254	Threonine	Glucose	0.82	648	Galactose	Threonic acid	0.84
255	Trigonelline	Glucose	0.67	649	Glucose	Threonic acid	0.82
256	Methylsuccinic acid	Glucose	0.89	650	Glyceric acid	Threonic acid	0.96
257	3MG	Glucose	0.87	651	Glycine	Threonic acid	0.77
258	beta-Alanine	GABA	0.72	652	Aspartic acid	Threonic acid	0.72
259	Fucose	GABA	0.89	653	Lactic acid	Threonic acid	0.70
260	Nicotinic acid	GABA	0.89	654	Linolenic acid	Threonic acid	0.86
261	Oxalic acid	GABA	0.79	655	L-Norleucine	Threonic acid	0.76
262	Ribitol	Glyceric acid	0.86	656	L-Phenylalanine	Threonic acid	0.73
263	Arabinose	Glyceric acid	0.93	657	Malic acid	Threonic acid	0.76
264	Citric acid	Glyceric acid	0.76	658	NAG	Threonic acid	0.78
265	D-Ribose	Glyceric acid	0.65	659	Palmitic acid	Threonic acid	0.72
266	Erythritol	Glyceric acid	0.78	660	Proline	Threonic acid	0.75
267	Ethanolamine	Glyceric acid	0.84	661	Urea	Threonic acid	0.69
268	Galactose	Glyceric acid	0.85	662	Threitol	Threonic acid	0.66
269	Glucose	Glyceric acid	0.85	663	Threonine	Threonic acid	0.96
270	Glycine	Glyceric acid	0.86	664	Methylsuccinic acid	Threonic acid	0.76
271	Aspartic acid	Glyceric acid	0.82	665	3MG	Threonic acid	0.78
272	Lactic acid	Glyceric acid	0.76	666	Ribitol	Threonine	0.89
273	Linolenic acid	Glyceric acid	0.88	667	Arabinose	Threonine	0.95
274	L-Norleucine	Glyceric acid	0.84	668	Citric acid	Threonine	0.69
275	Malic acid	Glyceric acid	0.85	669	Erythritol	Threonine	0.71
276	NAG	Glyceric acid	0.74	670	Ethanolamine	Threonine	0.73
277	Palmitic acid	Glyceric acid	0.79	671	Galactose	Threonine	0.84
278	Proline	Glyceric acid	0.83	672	Glucose	Threonine	0.82
279	Pyroglutamic acid	Glyceric acid	0.74	673	Glyceric acid	Threonine	0.94
280	Urea	Glyceric acid	0.80	674	Glycine	Threonine	0.78
281	Threitol	Glyceric acid	0.74	675	Aspartic acid	Threonine	0.72
282	Threonic acid	Glyceric acid	0.96	676	Lactic acid	Threonine	0.68
283	Threonine	Glyceric acid	0.94	677	Linolenic acid	Threonine	0.87
284	Methylsuccinic acid	Glyceric acid	0.86	678	L-Norleucine	Threonine	0.82
285	3MG	Glyceric acid	0.86	679	L-Phenylalanine	Threonine	0.66
286	2-HB	Glycine	0.80	680	Malic acid	Threonine	0.76
287	Arabinose	Glycine	0.73	681	NAG	Threonine	0.76
288	beta-Alanine	Glycine	0.69	682	Palmitic acid	Threonine	0.73
289	Citric acid	Glycine	0.90	683	Proline	Threonine	0.82
290	D-Ribose	Glycine	0.70	684	Urea	Threonine	0.68
291	Erythritol	Glycine	0.85	685	Threonic acid	Threonine	0.96
292	Ethanolamine	Glycine	0.92	686	Methylsuccinic acid	Threonine	0.77
293	Galactose	Glycine	0.81	687	3MG	Threonine	0.76
294	Glucose	Glycine	0.82	688	2-HB	Trigonelline	0.67
295	Glyceric acid	Glycine	0.86	689	beta-Alanine	Trigonelline	0.89
296	Aspartic acid	Glycine	0.80	690	Citric acid	Trigonelline	0.88
297	Lactic acid	Glycine	0.84	691	D-Ribose	Trigonelline	0.68
298	Linolenic acid	Glycine	0.83	692	Erythritol	Trigonelline	0.75
299	L-Norleucine	Glycine	0.92	693	Ethanolamine	Trigonelline	0.90
300	Malic acid	Glycine	0.97	694	Fucose	Trigonelline	0.71
301	NAG	Glycine	0.69	695	Galactose	Trigonelline	0.66
302	Palmitic acid	Glycine	0.87	696	Glucose	Trigonelline	0.67
303	Proline	Glycine	0.90	697	Glycine	Trigonelline	0.79
304	Pyroglutamic acid	Glycine	0.91	698	Lactic acid	Trigonelline	0.72
305	Quinic acid	Glycine	0.73	699	Linolenic acid	Trigonelline	0.67
306	Urea	Glycine	0.88	700	L-Norleucine	Trigonelline	0.80
307	Threitol	Glycine	0.78	701	Malic acid	Trigonelline	0.84
308	Threonic acid	Glycine	0.77	702	Palmitic acid	Trigonelline	0.75
309	Threonine	Glycine	0.78	703	Proline	Trigonelline	0.78
310	Trigonelline	Glycine	0.79	704	Pyroglutamic acid	Trigonelline	0.90
311	Methylsuccinic acid	Glycine	0.93	705	Quinic acid	Trigonelline	0.79
312	3MG	Glycine	0.96	706	Urea	Trigonelline	0.75
313	Ribitol	Aspartic acid	0.69	707	Threitol	Trigonelline	0.80
314	2-HB	Aspartic acid	0.77	708	Methylsuccinic acid	Trigonelline	0.81
315	Arabinose	Aspartic acid	0.69	709	3MG	Trigonelline	0.81
316	Citric acid	Aspartic acid	0.77	710	Ribitol	Methylsuccinic acid	0.66
317	Erythritol	Aspartic acid	0.74	711	2-HB	Methylsuccinic acid	0.84
318	Ethanolamine	Aspartic acid	0.79	712	Arabinose	Methylsuccinic acid	0.75
319	Galactose	Aspartic acid	0.74	713	beta-Alanine	Methylsuccinic acid	0.74
320	Glucose	Aspartic acid	0.76	714	Citric acid	Methylsuccinic acid	0.86
321	Glyceric acid	Aspartic acid	0.82	715	D-Ribose	Methylsuccinic acid	0.80
322	Glycine	Aspartic acid	0.80	716	Erythritol	Methylsuccinic acid	0.94
323	Lactic acid	Aspartic acid	0.69	717	Ethanolamine	Methylsuccinic acid	0.95
324	Linolenic acid	Aspartic acid	0.73	718	Galactose	Methylsuccinic acid	0.86
325	L-Norleucine	Aspartic acid	0.74	719	Glucose	Methylsuccinic acid	0.89
326	Malic acid	Aspartic acid	0.79	720	Glyceric acid	Methylsuccinic acid	0.86
327	NAG	Aspartic acid	0.67	721	Glycine	Methylsuccinic acid	0.93
328	Palmitic acid	Aspartic acid	0.81	722	Aspartic acid	Methylsuccinic acid	0.81
329	Proline	Aspartic acid	0.73	723	Lactic acid	Methylsuccinic acid	0.91
330	Pyroglutamic acid	Aspartic acid	0.71	724	Linolenic acid	Methylsuccinic acid	0.87
331	Quinic acid	Aspartic acid	0.76	725	L-Norleucine	Methylsuccinic acid	0.94
332	Urea	Aspartic acid	0.73	726	Malic acid	Methylsuccinic acid	0.96
333	Threitol	Aspartic acid	0.71	727	NAG	Methylsuccinic acid	0.74
334	Threonic acid	Aspartic acid	0.72	728	Palmitic acid	Methylsuccinic acid	0.93
335	Threonine	Aspartic acid	0.72	729	Proline	Methylsuccinic acid	0.93
336	Methylsuccinic acid	Aspartic acid	0.81	730	Pyroglutamic acid	Methylsuccinic acid	0.85
337	3MG	Aspartic acid	0.79	731	Quinic acid	Methylsuccinic acid	0.77
338	3-HBA	Inositol	0.68	732	Urea	Methylsuccinic acid	0.95
339	Isocitric acid	Inositol	0.80	733	Threitol	Methylsuccinic acid	0.88
340	Pipecolic acid	Inositol	0.66	734	Threonic acid	Methylsuccinic acid	0.76
341	3-HBA	Isocitric acid	0.84	735	Threonine	Methylsuccinic acid	0.77
342	Inositol	Isocitric acid	0.80	736	Trigonelline	Methylsuccinic acid	0.81
343	Pipecolic acid	Isocitric acid	0.80	737	3MG	Methylsuccinic acid	0.98
344	2-HB	Lactic acid	0.83	738	2-HB	3MG	0.82
345	Arabinose	Lactic acid	0.70	739	Arabinose	3MG	0.74
346	Citric acid	Lactic acid	0.68	740	beta-Alanine	3MG	0.73
347	D-Ribose	Lactic acid	0.80	741	Citric acid	3MG	0.89
348	Erythritol	Lactic acid	0.94	742	D-Ribose	3MG	0.78
349	Ethanolamine	Lactic acid	0.85	743	Erythritol	3MG	0.90
350	Galactose	Lactic acid	0.85	744	Ethanolamine	3MG	0.95
351	Glucose	Lactic acid	0.86	745	Galactose	3MG	0.85
352	Glyceric acid	Lactic acid	0.76	746	Glucose	3MG	0.87
353	Glycine	Lactic acid	0.84	747	Glyceric acid	3MG	0.86
354	Aspartic acid	Lactic acid	0.69	748	Glycine	3MG	0.96
355	Linolenic acid	Lactic acid	0.84	749	Aspartic acid	3MG	0.79
356	L-Norleucine	Lactic acid	0.84	750	Lactic acid	3MG	0.88
357	Malic acid	Lactic acid	0.88	751	Linolenic acid	3MG	0.86
358	NAG	Lactic acid	0.75	752	L-Norleucine	3MG	0.93
359	Palmitic acid	Lactic acid	0.83	753	Malic acid	3MG	0.98
360	Proline	Lactic acid	0.82	754	NAG	3MG	0.73
361	Pyroglutamic acid	Lactic acid	0.72	755	Palmitic acid	3MG	0.92
362	Urea	Lactic acid	0.93	756	Proline	3MG	0.92
363	Threitol	Lactic acid	0.86	757	Pyroglutamic acid	3MG	0.89
364	Threonic acid	Lactic acid	0.70	758	Quinic acid	3MG	0.75
365	Threonine	Lactic acid	0.68	759	Urea	3MG	0.92
366	Trigonelline	Lactic acid	0.72	760	Threitol	3MG	0.86
367	Methylsuccinic acid	Lactic acid	0.91	761	Threonic acid	3MG	0.78
368	3MG	Lactic acid	0.88	762	Threonine	3MG	0.76
369	Ribitol	Linolenic acid	0.81	763	Trigonelline	3MG	0.81
370	Arabinose	Linolenic acid	0.86	764	Methylsuccinic acid	3MG	0.98
371	Citric acid	Linolenic acid	0.78
372	D-Ribose	Linolenic acid	0.85
373	Erythritol	Linolenic acid	0.88
374	Ethanolamine	Linolenic acid	0.79
375	Galactose	Linolenic acid	0.99
376	Glucose	Linolenic acid	0.98
377	Glyceric acid	Linolenic acid	0.88
378	Glycine	Linolenic acid	0.83
379	Aspartic acid	Linolenic acid	0.73
380	Lactic acid	Linolenic acid	0.84
381	L-Norleucine	Linolenic acid	0.85
382	Malic acid	Linolenic acid	0.87
383	N-Acetyl-D-glucosamine	Linolenic acid	0.92
384	Palmitic acid	Linolenic acid	0.92
385	Proline	Linolenic acid	0.82
386	Pyroglutamic acid	Linolenic acid	0.75
387	Quinic acid	Linolenic acid	0.72
388	Urea	Linolenic acid	0.78
389	Threitol	Linolenic acid	0.84
390	Threonic acid	Linolenic acid	0.86
391	Threonine	Linolenic acid	0.87
392	Trigonelline	Linolenic acid	0.67
393	Methylsuccinic acid	Linolenic acid	0.87
394	3MG	Linolenic acid	0.86

• r: correlation coefficient; 2-HB: 2-Hydroxybutyric acid; D-AnG: 3,6-Anhydro-D-galactose; 3-HBA: 3-Hydroxybenzoic acid; NAG: N-Acetyl-D-glucosamine; GABA: gamma-Aminobutyric acid

Subsequent pathway enrichment analysis indicated that the identified markers were involved in multiple KEGG pathways. However, the metabolic network of C. arabica remains incompletely characterized, with limited experimentally validation of enzymes and reactions. Although some pathways such as ABC transporters included several annotated metabolites, their biochemical roles in C. arabica remain unclear. In contrast, carbon metabolism and biosynthesis of amino acids are better characterized. Within carbon metabolism, four markers (glycine, glyceric acid, malic acid, and citric acid) were involved. Noticeably, malic acid and citric acid are biochemically linked through reaction sequences involving oxaloacetic acid as a key intermediate and aspartic acid associated with transamination reactions. The biosynthesis of amino acid pathway also included glycine, proline, citric acid and threonine, with particular emphasis on the reversible conversion between glycine and threonine. These findings highlight coordinated alterations in central carbon and amino acid metabolism that may underlie the metabolic differentiation between civet and unprocessed coffee beans.

Overall, the results of untargeted GC-MS-based metabolomic analysis were consistent with previous studies reporting differences in volatile compounds, such as caffeine, trigonelline, amino acids and sugars and phenolic constituents, between civet and normal coffee (Farag et al., 2023; Hájíček et al., 2025). The key discriminative metabolites identified here were primarily associated with sugars, organic acids, amino acids, and fatty acids, indicating substantial perturbations in central carbon and amino acid metabolism. Furthermore, correlation network and pathway enrichment analyses suggested a coordinated and system-wide metabolic shift, rather than isolated changes in individual compounds. Collectively, these findings highlight the profound impact of civet gastrointestinal processing on the chemical composition of coffee beans and provide a foundation for identifying metabolite-based quality markers for civet coffee authentication.

Two fatty acids, palmitic and linolenic acid, were identified among the twenty-three potential quality markers. In the present study, both compounds were downregulated in civet coffee. However, a previous study comparing Robusta civet coffee with normal coffee using GC-MS analysis of fatty acid methyl esters reported no significant differences in the levels of these two fatty acids between groups. This discrepancy may reflect species-specific metabolomic responses, suggesting that metabolite profiles are not conserved across coffee varieties and that metabolic pathway modifications in Robusta civet coffee may differ from those in Arabica. In addition, the limited sample size in the previous study (n = 5 per group) may have reduced its representativeness (Mitra et al., 2025).

Glyceric acid, malic acid, arabinose, and citric acid were also highlighted as noticeable markers for the civet coffee group, all exhibiting a decreasing trend. These compounds have previously been reported in GC-MS-based metabolomic analysis between civet and non-civet coffee (Jumhawan et al., 2015). In that study, the discriminative power of malic acid and citric acid was further supported by a subsequent qualitative validation. However, their peak intensities were reported to be upregulated in civet coffee, in contrast to the present findings. This inconsistency may be attributed to the small validation dataset (n = 3 per group), which limits the robustness and generalizability of the results.

It is suggested that the gastrointestinal fluid with numerous digestive enzymes and microbiome induces the metabolic shift of coffee beans. Iswanto et al. (2019) identified Methylobacterium populi, Klebsiella quasipneumoniae, Raoultella ornithinolytica and Stenotrophomonas chelatiphaga in Luwak feces eating green coffee bean. In addition, Watanabe et al. (2020) reported that Gluconobacter species, such as G. frateurii and G. japonicus, may dominate the microbiome and influence the tricarboxylic acid (TCA) cycle of coffee beans. Notably, Gluconobacter spp. lack succinate dehydrogenase, resulting in an incomplete TCA cycle, which has been proposed to promote the accumulation of malic and citric acids in civet coffee (Watanabe et al., 2020). However, the opposite trend observed in the present study may reflect geographical variation in civet gut microbiome composition. Furthermore, as malic acid and citric acid are involved in multiple KEGG pathways, their downregulation cannot be attributed to a single metabolic mechanism.

Differences in civet diet between Indonesian and Vietnamese production systems may partly explain the inconsistent organic acid profiles reported across studies. Diet strongly influences gut microbiota composition and microbial fermentation activity (Zhang, 2022). In Indonesian kopi luwak production, civets are commonly fed mainly coffee cherries and fruits, whereas the civets in the present Vietnamese farm received mixed diets including rice, vegetables, meat, and coffee cherries. Such dietary variation may alter microbial communities and substrate availability in the gastrointestinal tract, thereby influencing tricarboxylic acid-related metabolism and contributing to the observed downregulation of malic and citric acids in our civet coffee samples compared with previous reports (Jia et al., 2025; Li et al., 2025).

This study provides preliminary insights into the metabolomic profile of civet coffee and proposes candidate markers for authentication. However, several limitations should be acknowledged. First, the untargeted GC-MS approach provides only relative quantification, and further targeted validation using larger and more diverse sample sets is required. Although the sample size used here is acceptable for exploratory metabolomic studies, broader sampling across geographic regions and civet populations would improve the robustness and generalizability of the findings. Second, volatile compounds, which contribute substantially to coffee aroma, were not comprehensively characterized. Third, despite derivatization, GC-MS has limited metabolome coverage, and the integration of high-resolution LC-HRMS would improve the detection and structural elucidation of polar and semi-polar metabolites, including alkaloids, peptides, and polyphenols. Finally, civet gut microbiota likely plays a central role in shaping metabolomic variation, but microbiome composition was not directly characterized in this study. Future studies integrating metabolomics, microbiome profiling, and sensory analysis may therefore provide a more comprehensive framework for civet coffee authentication and quality assessment.

Conclusion

In this study, GC-MS-based metabolomics was used to profile civet and unprocessed green coffee beans and identify quality markers. Multivariate analysis showed clear group separation, indicating a distinct civet-associated metabolomic signature. Twenty-three metabolites were proposed as markers, with mannitol consistently upregulated, while most sugars, organic acids, and amino acid-related compounds were downregulated. These changes suggest a coordinated metabolic shift driven by digestive processes, though further validation with targeted and high-resolution methods is required due to limited metabolome coverage.

Author Contribution

Conceptual idea: Hoa, Q. N.; Oanh, T. K. N.; Methodology design: Nhung, P. N.; Oanh, T. K. N.; Data collection: Hai, D. L.; Nguyen, T. L.; Giang, N. T.; Tam, T. T. T.; Data analysis and interpretation: Nhung, P. N.; Quang, H. V. P.; and Writing and editing: Nhung, P. N.; Hoa, Q. N.; Oanh, T. K. N.

Acknowledgments

This work was supported by the University of Science and Technology of Hanoi (USTH) under grant number USTH.GED.01/24-25 .

Data Availability Statement

Data available upon request to authors.

References

• Chan, S., & Garcia, E. (2011). Comparative physicochemical analyses of regular and civet coffee. The Manila Journal of Science, 7(1):19-23.
• Diez-Simon, C., Mumm, R., & Hall, R. D. (2019). Mass spectrometry-based metabolomics of volatiles as a new tool for understanding aroma and flavour chemistry in processed food products. Metabolomics, 15:41.
• Erban, A. et al. (2019). Discovery of food identity markers by metabolomics and machine learning technology. Scientific Reports, 9:9697.
• Farag, M. A. et al. (2023). Metabolite profiling of premium civet luwak bio-transformed coffee compared with conventional coffee types, as analyzed using chemometric tools. Metabolites, 13(2):173.
• Farah, A. et al. (2005). Effect of roasting on the formation of chlorogenic acid lactones in coffee. Journal of agricultural and food chemistry, 53(5):1505-1513.
• Frolova, N. et al. (2026). Gas chromatography-mass spectrometry (GC-MS) in the plant metabolomics toolbox: GC-MS in multi-platform metabolomics and integrated multi-omics research. International Journal of Molecular Sciences, 27(3):1343.
• Garret, R. et al. (2014). Ambient mass spectrometry employed for direct analysis of intact arabica coffee beans. Journal of the Brazilian Chemical Society, 25(7):1172-1177.
• Gigl, M. et al. (2021). Identification and quantitation of reaction products from quinic acid, quinic acid lactone, and chlorogenic acid with strecker aldehydes in roasted coffee. Journal of agricultural and food chemistry , 69(3):1027-1038.
• Ginz, M., & Engelhardt, U. H. (2001). Identification of new diketopiperazines in roasted coffee. European Food Research and Technology, 213:8-11.
• Haile, M., & Kang, W. H. (2019). The role of microbes in coffee fermentation and their impact on coffee quality. Journal of Food Quality, 2019:4836709.
• Hájíček, J. et al. (2025). Comparative analysis of chemical composition and antioxidant activity in conventional, civet, and elephant coffees: Is there a definitive authentication marker of elephant coffee? Beverages, 11(3):79.
• Halket, J. M. et al. (2005). Chemical derivatization and mass spectral libraries in metabolic profiling by GC/MS and LC/MS/MS. Journal of Experimental Botany, 56(410):219-243.
• Ifmalinda, I. et al. (2019). Chemical characteristics comparison of palm civet coffee (kopi luwak) and arabica coffee beans. Journal of Applied Agricultural Science and Technology, 3(2):280-288.
• Iswanto, T. et al. (2019). Isolation and identification of caffeine-degrading bacteria from soil, coffee pulp waste and excreted coffee bean in Luwak feces. Biodiversitas Journal of Biological Diversity, 20(6):1580-1587.
• Jia, Y. et al. (2025). From diet to microbiota: How fruits and vegetables influence gut microbiota. International Journal of Food Science and Technology, 60(1):vvae008, 2025.
• Jumhawan, U. et al. (2015). Application of gas chromatography/flame ionization detector-based metabolite fingerprinting for authentication of Asian palm civet coffee (Kopi Luwak). Journal of Bioscience and Bioengineering, 120(5):555-561.
• Jumhawan, U. et al. (2013). Selection of discriminant markers for authentication of Asian palm civet coffee (Kopi Luwak): A metabolomics approach. Journal of agricultural and food chemistry , 61(33):7994-8001.
• Li, Y. et al. (2025). Formation mechanism and nutritional effects of organic acids in fermented foods: A review. Journal of Food Science, 90(12):e70745.
• Marcone, M. F. (2004). Composition and properties of Indonesian palm civet coffee (Kopi Luwak) and Ethiopian civet coffee. Food Research International, 37(9):901-912.
• Mitra, R. et al. (2025). Civet Robusta and natural Robusta coffee are different on key fatty acid methyl esters and total fat. Scientific Reports , 15:36281.
• Muzaifa, M. et al. (2020). Chemical composition of green and roasted coffee bean of Gayo arabica civet coffee (kopi luwak). Paper presented at the IOP Conference Series: Earth and Environmental Science, 425:012001.
• Ollennu-Chuasam, P. et al. (2025). Effect of the roast level on lipophilic and hydrophilic compounds in pot and filter coffee beverages. ACS Food Science & Technology, 5(6):2187-2202.
• Pereira, G. V. M. et al. (2025). Integrated microbial-metabolomic analysis reveals how fermentation contributes to the unique flavor of African Arabica coffee. Food Chemistry: Molecular Sciences, 12:100344.
• Pilipczuk, T. et al. (2015). The influence of roasting and additional processing on the content of bioactive components in special purpose coffees. Journal of Food Science and Technology, 52:5736-5744.
• Selamat, J., Rozani, N. A. A., & Murugesu, S. (2021). Application of the metabolomics approach in food authentication. Molecules, 26(24):7565.
• Şen, D., & Gökmen, V. (2022). Kinetic modeling of Maillard and caramelization reactions in sucrose-rich and low moisture foods applied for roasted nuts and seeds. Food Chemistry, 395:133583.
• R Core Team. (2021). R: A language and environment for statistical computing.
• Takeda, H. et al. (2024). MS-DIAL 5 multimodal mass spectrometry data mining unveils lipidome complexities. Nature Communications, 15:9903.
• Thévenot, E. A. et al. (2015). Analysis of the human adult urinary metabolome variations with age, body mass index, and gender by implementing a comprehensive workflow for univariate and OPLS statistical analyses. Journal of Proteome Research, 14(8):3322-3335.
• Watanabe, H. et al. (2020). Gluconobacter dominates the gut microbiome of the Asian palm civet Paradoxurus hermaphroditus that produces kopi luwak. PeerJ, 8:e9579.
• Zayed, A. et al. (2022). Dissecting coffee seeds metabolome in context of genotype, roasting degree, and blending in the Middle East using NMR and GC/MS techniques. Food Chemistry , 373:131452.
• Zhang, P. (2022). Influence of foods and nutrition on the gut microbiome and implications for intestinal health. International Journal of Molecular Sciences, 23(17):9588.
• Zhong, P. et al. (2022). Untargeted metabolomics by liquid chromatography-mass spectrometry for food authentication: A review. Comprehensive Reviews in Food Science and Food Safety, 21(3):2455-2488.

• Editor de seção:
  Renato Paivahttp://orcid.org/0000-0001-5107-0285

Publication Dates

• Publication in this collection
  07 Sept 2026
• Date of issue
  2026

History

• Received
  20 Apr 2026
• Accepted
  17 June 2026