Raman-Chemometric Framework for Rapid Authentication of Edible Oils in Processed Foods Using a Validated One Step Sampling Technique

Raman-Chemometric Framework for Rapid Authentication of Edible Oils in Processed Foods Using a Validated One Step Sampling Technique | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Raman-Chemometric Framework for Rapid Authentication of Edible Oils in Processed Foods Using a Validated One Step Sampling Technique Amrita Shaw, Chandrasekar SUBRAMANI NARAYANA, Sai Muthukumar V, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8872116/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 9 You are reading this latest preprint version Abstract Authentication of cooking oils in processed foods is essential for food safety, quality control, and regulatory compliance, yet routine analysis remains constrained by solvent-intensive extraction procedures and limited applicability to complex food matrices. This study presents a proof-of-concept Raman–chemometric framework for direct oil authentication in processed foods using minimal, solvent-free sample handling. Potato chips were selected as a representative fried food matrix, and five commonly used edible oils (sunflower, soybean, groundnut, palm, and vanaspati) were analyzed in both pure form and corresponding chip matrices to enable systematic cross-matrix evaluation. A one-step tissue paper blotting method was employed for rapid oil recovery, followed by Raman spectroscopic analysis. Fatty-acid-associated Raman bands were identified and systematically combined into chemically interpretable inter-peak intensity ratios reflecting variations in saturation and ester content. A statistically grounded two-stage marker selection workflow integrating Random Forest importance ranking with non-parametric Kruskal–Wallis testing was applied, followed by using one-way analysis of variance and coefficient of variation analysis. Five robust ratiometric markers (I₁₆₅₂/₁₇₄₂, I₁₆₅₂/₁₄₃₄, I₁₇₄₂/₁₂₅₉, I₁₆₅₂/₁₂₅₉, and I₁₂₉₆/₁₄₃₄) showed strong association with saturated versus unsaturated fatty acid profiles. Multivariate analysis based on these markers revealed pronounced separation among oil types (91% explained variance; F-values up to 22,629; p < 0.001) and statistically significant discrimination within chip matrices (77% explained variance), despite attenuation effects from the food matrix. Multivariate analysis of variance confirmed robust separation (Wilks’ Λ = 0.0307; p < 0.0001). Overall, this framework establishes an interpretable, extraction-free, and scalable foundation for high-throughput oil authenticity screening in processed foods. Raman spectroscopy Edible oil authentication Ratiometric analysis PCA ANOVA MANOVA Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Introduction Identifying the type of oil used in processed foods is crucial due to its significant impact on consumer health (Foster et al. 2009 ; Moore et al. 2012 ). A common form of adulteration involves substituting listed oils with cheaper analogues, locally available alternatives, or repeatedly heated oils, compromising both food quality and safety (Rifna et al. 2022 ; Haji et al. 2023 ). Some of these substitutes, such as argemone oil, contain toxins and are banned globally by regulatory agencies (Tan et al. 2021 ; Begum and Jain 2024 ). Even legally permitted oils like palm oil and vanaspati pose health risks and are unsuitable for individuals with cardiovascular diseases and various health issues, while soybean and groundnut oils can trigger allergic reactions (Johnson and Saikia 2009 ; Dorni et al. 2018 ). Although oil adulteration is a global issue, it becomes a formidable challenge in developing countries like India (Rana et al. 2014 ). Regulatory enforcement mainly targets the organized food sector, leaving much of the unregulated market vulnerable. As a result, incidents of such adulteration have increased significantly in recent years (Program 2024 ). Moreover, recent investigations suggest that economically motivated adulteration is no longer confined to unregistered businesses. A nationwide audit by the Food Safety and Standards Authority of India (FSSAI) in 2022, seized thousands of liters of edible oils for suspected mislabeling, fake branding and exceeding levels of trans-fatty acids, highlighting the urgent need for stricter monitoring and legal intervention (Sharma R, 2022 ; Desk, 2024 ). Typically, Chromatographic methods like High-Performance Liquid Chromatography (HPLC) (Cercaci et al. 2003 ; Navratilova et al. 2022 ; Khursheed 2024 ) and Gas Chromatography (GC) (Aparicio and Aparicio-Ruı́z 2000 ; Su et al. 2011 ; Ün and OK 2018 ), along with Nuclear Magnetic Resonance (NMR) (Ün and OK 2018 ; Gu et al. 2021 ), are used for oil authentication. While these methods provide high accuracy, their complex sample processing protocols, lengthy analysis procedures and high operational costs make them unsuitable for rapid, real time monitoring (Elumalai and Natarajan 2025 ). Thus, a high-throughput screening technique with minimal sample processing is essential and a desirable alternative. Vibrational spectroscopic techniques have emerged as a superior alternative to chromatographic methods, overcoming their limitations and enabling real-time monitoring of edible oils (Yadav 2018 ; Hu et al. 2019 ; Yao et al. 2020 ; Forooghi 2023 ). Among these, Raman spectroscopy stands out for its ability to provide detailed molecular fingerprints via Raman scattering (Mendes et al. 2015 ; Yang et al. 2019 ; Forooghi 2023 ). It offers a promising alternative because it directly probes molecular vibrations associated with lipid structure, including saturation, unsaturation and ester functionality (Baeten et al. 1998 ; Huang et al. 2016 ; Kwofie et al. 2020 ; Wang et al. 2025 ). Characteristic Raman bands at critical wavenumbers like 1650–1660 cm⁻¹ (C = C), 1740–1750 cm⁻¹ (ester carbonyl), and 1430–1440 cm⁻¹ (CH₂ scissoring) have been widely used to differentiate oil types and monitor oxidation (Huang et al. 2022 ; Wang et al. 2025 ). Attempts have been made to probe the quality of waste cooking oils (Huang et al. 2016 ) and the quantification of trans-fat in edible oils (Gong et al. 2019 ) using Raman spectroscopy. Therefore, the current trends in food quality monitoring clearly indicate the adaption of molecular spectroscopic techniques like Raman Spectrometry for the development of rapid quality monitoring tools. Ratiometric analysis in spectroscopy involves calculating the ratio of intensities between two characteristic peaks rather than using absolute intensity values, that enhances measurement sensitivity and selectivity (Wu et al. 2024 ). Ratiometric interpretation of Raman bands is increasingly common due to its robustness against intensity fluctuations, laser defocus, and sample variability (Jamieson et al. 2018). However, the literature lacks ratiometric markers that are systematically derived, reproducible, statistically validated, and correlated with chemical composition. Stringent selection of these ratiometric markers helps to reduce the high dimensionality of spectral data and improve statistical performance (Jamieson et al. 2018). Unfortunately, most existing studies rely only on visual or empirical selection of ratios without rigorous statistical validation. Therefore, it is important to develop a systematic framework for identifying and validating ratiometric markers through quantitative correlation analysis and multivariate statistical testing. The application of multivariate analysis to spectral data highlights the growing importance of data analysis in food quality monitoring (Vaskova and Buckova 2018 ). Chemometric analysis combined with spectrometry has been widely recognized for its effectiveness, particularly in oil quality assessment (Gómez-Caravaca et al. 2016 ). Integrating vibrational spectroscopic techniques with chemometric methods is expected to further enhance the evaluation of edible oil quality (Liu et al. 2020 ). Rigorous application of chemometrics for oil authentication in complex food matrices is gap in the scientific domain which demands more efforts. Quality monitoring in processed food is more challenging than quality monitoring in ingredients due to matrix interference. One such example is interference from starch and seasonings while screening for the oil in potato chips (food matrix). Most of the food quality monitoring studies deal with ingredients testing rather than testing the final food product (Inteaz 1989 ) as testing of such end products involves elaborate sampling, extraction, isolation, and derivatization. Studies have shown that crucial components, such as fats and oils, which are extensively used in food preparation still need to be isolated from matrices (Bansal et al. 2010 ) via a long, tedious process before impending analysis (Kim et al. 2018 ; Yao et al. 2020 ; Kaimal et al. 2021 ). As per example, the traditional method for oil analysis in fried food matrix involves mechanical pressing to extract oil followed by debris removal, filtration and additional purification steps (Aykas and Rodriguez-Saona 2016 ). This extended processing time underscores the need for developing a rapid, high throughput, inexpensive and efficient sampling protocol. As seen above, most of the reported studies have explored oil type analysis using Raman-PCA (Castro et al. 2022 ) involving tedious traditional multi-step processes. Little efforts have been made on developing techniques with minimal sample processing, directly from the food matrix, which can actually make the oil quality monitoring highly reliable and reproducible. From a food safety perspective, a framework capable of analyzing oil types from food matrices is inevitably crucial. In this project, we have developed Raman-PCA frameworks to identify oil types in processed foods (Potato chips as the food matrix) with minimal sample processing. A novel, non-invasive tissue paper-based oil sampling technique for food matrix was developed and validated in the study. Potato chips fried in five different edible oils, including sunflower oil, vanaspati oil, palm oil, groundnut oil, and soybean oil, were explored to design the framework. While these oils are not always classified as adulterants, they can pose health risks to particular individuals with soy or peanut allergies or cardiovascular conditions when added undeclared. For developing the framework, spectral analysis identified fatty acid-associated Raman bands from which inter-peak intensity ratios were systematically constructed. A two-stage selection strategy combining Random Forest marker importance ranking and Kruskal-Wallis statistical testing identified the most discriminatory intensity ratios, which were further validated through one-way ANOVA and coefficient of variance (CV) analysis for marker reproducibility and robustness. The selected ratiometric markers were subjected to multivariate analysis, with principal component analysis (PCA) performed to assess class separability, followed by formal statistical validation using MANOVA to confirm the robustness of multivariate separation beyond visual inspection. Materials and methods Procurement and preparation of samples Five commercially available packaged edible oils (Sunflower-SO, Soybean-SOYO, Groundnut-GNO, Palm-PO, and Vanaspati-VO) with FSSAI certification were procured from retail outlets in Anantapur district, Andhra Pradesh, India. Figure 1 depicts the steps employed for SO, which was systematically replicated across all oil types. The samples, along with their codes, are provided in Table 1 . To mimic actual practices followed by small scale potato chips manufacturing units, thin slices of potatoes were fried in each oil over 9 successive batches, with systematic collection of both fried chips and oil samples. Table 1 List of samples prepared for analysis Sample Codes Heating Stages Sunflower Oil (SO) Palm Oil (PO) Vanaspati (VO) Soya Bean Oil (SOYO) Groundnut Oil (GNO) Oil Food matrix (Chips) Oil Food matrix (Chips) Oil Food matrix (Chips) Oil Food matrix (Chips) Oil Food matrix (Chips) 0 SO 0 - PO 0 - VO 0 - SOYO 0 - GNO 0 - 1 SO 1 SO C 1 PO 1 PO C 1 VO 1 VO C 1 SOYO 1 SOYO C 1 GNO 1 GNO C 1 2 SO 2 SO C 2 PO 2 PO C 2 VO 2 VO C 2 SOYO 2 SOYO C 2 GNO 2 GNO C 2 3 SO 3 SO C 3 PO 3 PO C 3 VO 3 VO C 3 SOYO 3 SOYO C 3 GNO 3 GNO C 3 4 SO 4 SO C 4 PO 4 PO C 4 VO 4 VO C 4 SOYO 4 SOYO C 4 GNO 4 GNO C 4 5 SO 5 SO C 5 PO 5 PO C 5 VO 5 VO C 5 SOYO 5 SOYO C 5 GNO 5 GNO C 5 6 SO 6 SO C 6 PO 6 PO C 6 VO 6 VO C 6 SOYO 6 SOYO C 6 GNO 6 GNO C 6 7 SO 7 SO C 7 PO 7 PO C 7 VO 7 VO C 7 SOYO 7 SOYO C 7 GNO 7 GNO C 7 8 SO 8 SO C 8 PO 8 PO C 8 VO 8 VO C 8 SOYO 8 SOYO C 8 GNO 8 GNO C 8 9 SO 9 SO C 9 PO 9 PO C 9 VO 9 VO C 9 SOYO 9 SOYO C 9 GNO 9 GNO C 9 Tissue-paper based oil sampling for food matrix (potato chips) A novel, simplistic but promising tissue paper–based sampling technique tailored for food matrices was developed and evaluated in this study, with potato chips used as a representative model system as a food matrix. The proposed extraction approach is intended as a rapid and facile screening tool for oil authentication. Single-ply, laboratory-grade, non-fragrant tissue paper (Kimwipes®, Kimberly-Clark) was cut into uniform sections and placed between two fried potato chips with consistent pressure to facilitate oil absorption through capillary action. This method enabled efficient transfer of surface oil into the tissue paper via blotting. These tissue papers which assimilated oil from the fried chips were subsequently analyzed directly by Raman spectroscopy without any further processing. A pictorial depiction of the developed extraction technique is presented in Fig. 2 . Data acquisition Raman spectroscopic measurements of all oil and chip samples were performed using a dispersive Raman microscope system (DXR, Thermo Fisher Scientific) equipped with a 785 nm excitation laser. After preliminary spectral acquisition survey, the instrumental parameters were optimized as follows: laser power, 20 mW; exposure time, 30 s; number of exposures, 2; confocal aperture, 25 µm slit; diffraction grating, 400 lines mm⁻¹; spectral resolution, 4 cm⁻¹. Raman spectra were acquired over the typical molecular finger printing range of 200–3000 cm⁻¹. To ensure statistical reliability and account for sample heterogeneity, 20 different spectra were collected by focusing on different locations on each sample. This spatially resolved acquisition strategy enhances measurement robustness and reduces local sampling bias. The resulting spectra were treated as technical replicates representative of each experimental condition, and all subsequent analyses were performed accordingly under controlled laboratory settings. The complete dataset comprised 1000 Raman spectra from oil samples (200 spectra per oil type across 10 heating stages, including pristine) and 900 spectra from chip samples (180 spectra per oil type across 9 frying stages, starting from 1st frying as chips cannot exist in pristine form), totaling 1900 spectra across five oil types. Data preprocessing The spectral dataset was subjected to a systematic preprocessing pipeline comprising three sequential steps (Fig. 3). (i) Baseline correction was performed using the asymmetric least squares (ALS) (Korepanov 2020 ) algorithm with optimized parameters: asymmetry parameter (p) = 0.01, smoothing parameter (λ) = 10⁵, and 10 iterations. (ii) Spectral smoothing was applied using the Savitzky–Golay filter (Barton et al. 2018 ) with an 11-point window and third-order polynomial fitting to reduce noise while preserving spectral features. (iii) Finally, the entire dataset was normalized to the 2720 cm⁻¹ band, which corresponds to C–H stretching vibrations found commonly in all fatty acid chains across all samples. Specifically, 2720 cm − 1 was chosen as the wavenumber for normalization as it will not hamper the relative intensities of other functionally important peaks (Baeten et al. 1998 ; Kwofie et al. 2020 ). Figure 3 Schematic flow for spectral data processing used in the proposed framework Construction of ratiometric markers The key component of identifying and differentiating various oil in the proposed framework is attempted by Ratiometric analysis. For ratiometric analysis, spectral peaks corresponding to the following major vibrational modes of fatty acids and esters were systematically identified; for example, Carbonyl stretching of carboxyl groups (–COOH) vibration at 1742 cm⁻¹, methylene bending (δ-CH₂) at 1434 cm⁻¹, C-C skeletal vibrations at 1123 cm⁻¹ and 1296 cm⁻¹, C-H stretching at 2720 cm⁻¹, and -CH 2 stretching at 3010 cm⁻¹. Unsaturated fatty acids exhibited additional characteristic bands at 1259 cm⁻¹ (= C–H bending vibrations) and 1652 cm⁻¹ (C = C stretching vibrations), with peak intensities directly proportional to the degree of unsaturation. All possible pairwise intensity ratios (Ratiometric markers) were computed using these key spectral bands. Optimal ratiometric markers were identified using statistical analysis and validation, as described below. Statistical selection, marker validation, and multivariate modeling Statistical identification of ratio markers Optimization and validation of the ratiometric markers were carried out following the analytical workflow shown in Fig. 4 . All analyses were performed using open source Python package 3.12.9. After constructing the comprehensive set of ratiometric markers, a two-stage selection strategy were employed to identify the most discriminatory markers. First, a Random Forest (RF) classifier was applied to rank all ratiometric markers based on their marker importance scores. Random Forest creates an ensemble of decision trees trained on randomized data subsets and assigns feature importance based on their contribution to classification accuracy across the forest (Menze et al. 2009 ; Shehata et al. 2024 ). The RF model provided an independent, model-based assessment of each marker. It should be noted that feature importance rankings obtained from Random Forest models may be influenced by feature correlations and dataset characteristics and should therefore be interpreted in conjunction with complementary statistical analyses. The markers exhibiting the highest importance scores were chosen as the most effective markers and selected for further statistical validation using the Kruskal-Wallis test (Kaur et al. 2016 ; Pereira et al. 2021 ), a non-parametric method for assessing group-wise differences. The ratios demonstrating statistically significant variations across oil classes (p < 0.05) in the Kruskal-Wallis analysis were selected for further screening and validation through one-way analysis of variance (ANOVA) (Davari et al. 2024 ) to confirm discriminatory power. This validation step was further aided by coefficient of variance (CV) analysis to assess marker reproducibility and robustness. Ultimately, Principal Component Analysis (PCA) (Jolliffe and Cadima 2016 ) was performed on the final set of the ratiometric markers thus obtained to assess their effectiveness in clustering different oil classes in multivariate space. The variance explained by each principal component were calculated, and score plots were generated to visualize clustering. To check whether the clustering observed in PCA is statistically significant, MANOVA (Zhu et al. 2022 ) was performed on the ratios. Four complementary test statistics were computed: Wilks' lambda (Λ), Pillai's trace (V), Hotelling–Lawley trace (T), and Roy's greatest root (Θ) ( Alkarkhi and Alqaraghuli 2019 ) to confirm the robustness of multivariate separation. The marker selection and validation procedures were designed to be systematic and statistically grounded within the available dataset and within the scope of the present proof-of-concept study. Nevertheless, further evaluation using independent validation datasets will be valuable for confirming long-term marker stability and generalizability under broader application scenarios. Results and discussion Interpretation of Raman spectral peaks related to fatty acids and esters Spectral analysis was performed on the baseline-corrected, smoothened and normalized Raman spectra of five oil types and their corresponding chip samples (Fig. 5). Both oils and chips samples were found to contain a few strong peaks at 1123 cm⁻¹ [(C–C) stretching of the (CH2) n group], 1259 cm − 1 (= C–H bending), 1296 cm⁻¹ [(= C–H) deformation of cis(R–HC = CH–R)], 1434 cm⁻¹ (methylene bending (δ-CH₂)), 1652 cm⁻¹ [(C = C) of cis(R-HC = CH-R)], 1742 cm⁻¹ [carbonyl stretching of ester carboxyl groups (-COOH)], 2720 cm⁻¹ (C–H stretching) and 3010 cm⁻¹ (methyl group stretching). As mentioned in materials and methods section, the chips spectra were collected after blotting the oil from the chips using tissue paper. Therefore, presence of similar signature peaks in chips and oils suggests that the dominant molecular signatures of the oils are largely preserved under the proposed sampling conditions, despite effects from both the food matrix and the tissue paper (cellulose) matrix. It suggests tissue paper based sampling to be an effective alternative for tedious extraction based sampling which are most commonly used during food ingredient analysis. Overall, the preliminary spectral analysis of the data acquired from the oil laden tissue paper confirmed that Raman spectroscopy captured chemically meaningful lipid signatures in both oil and chips, paving way for a robust physicochemical foundation for subsequent ratiometric analysis and multivariate discrimination. After assigning the peaks to specific vibrational groups, the peak intensities were systematically evaluated to determine their correlation with variations in saturated and unsaturated fatty acid composition across the studied oils. CV analysis was done on the intensity of the above mentioned peaks. Unfortunately, CV obtained, as low as 0.2–0.3%, indicated their incapability to be used as markers. In order to find suitable markers, instead of relying on the individual peak intensities, intensity ratios of two chemically meaningful peaks (ratiometric markers) were explored. Figure 5 Baseline corrected, smoothened and normalized Raman spectra collected from oils and chips Identifying ratiometric markers suitable for oil type classification To identify the suitable ratiometric markers from the extensive pool of all potential markers, three crucial criteria were evaluated; chemical relevance, model-based marker-ranking approach and statistical significance. Only those ratios were retained which (i) were constructed from chemically interpretable Raman bands, (ii) showed high significance in Random forest markers importance and (iii) statistically significant as per Kruskal–Wallis analysis. Ratios failing to meet these criteria were excluded. Random Forest (RF) classifier was employed on the chemically relevant markers as an independent, model-based marker-ranking approach. According to the RF model (Table 2 ), five ratiometric markers I₁₆₅₂/₁₇₄₂, I₁₆₅₂/₁₄₃₄, I₁₇₄₂/₁₂₅₉, I₁₆₅₂/₁₂₅₉, and I₁₂₉₆/₁₄₃₄ were identified as the most informative. Their importance values ranged from ~ 0.300 to 0.116, while all other ratios showed much lower contributions. The statistical significance (p < 0.005) of these highly ranked ratios were further assessed through Kruskal-Wallis group-wise testing, which revealed the same ratios to consistently show highly significant differences (p < < 10⁻⁶) across oil classes. This systematic selection resulted in a compact, chemically significant ratiometric marker set, which formed the basis for all subsequent multivariate classification. Table 2 Statistical Significance by Kruskal–Wallis analysis and Random Forest Importance of Ratiometric markers in Oils and Chips Ratio markers Kruskal–Wallis analysis RF importance Oils Chips Oils Chips p value p value I 1652/1434 p < 1 x 10 − 22 p < 1 x 10 − 15 0.300 0.417 I 1296/1434 p < 1 x 10 − 6 p < 1 x 10 − 9 0.116 0.139 I 1652/1742 p < 1 x 10 − 10 p < 1 x 10 − 8 0.208 0.138 I 1742/1259 p < 1 x 10 16 p < 1 x 10 − 13 0.212 0.131 I 1652/1259 p < 1 x 10 − 12 p < 1 x 10 − 16 0.123 0.103 Statistical screening and reproducibility validation of ratiometric markers A cross-validated marker selection strategy was employed to systematically validate the ratiometric markers identified in previous section. Based on the ANOVA results (F-statistic) and CV values (Table 3 ), all five ratiometric markers: I₁₆₅₂/I₁₄₃₄, I₁₇₄₂/I₁₂₅₉, and I₁₆₅₂/I₁₇₄₂, I 1652/1259 and I 1296/1434 were identified as suitable for optimal oil type classification. For oil samples, these ratios exhibited strong statistical separation across oil types (F: 22629.79, 1267.22, and 11697.20 and CV: 43.77%, 43.86% and 38.23%, respectively). However, for chips samples, the F-statistics and CV values turned out to be lower (F: 2568.23, 331.46 and 282.45; CV: 43.64%, 38.82% and 33.09%). This reduction is most likely attributable to the increased heterogeneity and compositional complexity of the chips (food matrix). Overall, the results from the ratio markers analysis demonstrated that oil type classification can be potentially achieved using this identified and statistically significant set of five robust, chemically interpretable ratio markers and suitable for chemometric analysis for the final framework development. Table 3 Statistical Significance of selective ratiometric markers assessed by ANOVA (F_statistic) and Coefficient of Variance (CV) values for oils and chips. Ratiometric Markers Oil Samples Chips Samples F_statistic CV Values (%) F_statistic CV Values (%) I 1652/1434 22629.79 43.77 2568.23 43.64 I 1652/1742 11697.20 43.86 331.46 38.82 I 1742/1259 1267.22 38.23 282.45 33.09 I 1652/1259 666.20 34.52 55.76 10.58 I 1296/1434 637.99 33.28 189.10 20.48 Multivariate classification and statistical validation using optimized ratiometric markers The classification efficiency of the selected ratio markers was further evaluated using multivariate analysis to assess whether oil types can form distinct clustering corresponding to different types of oils. Multivariate separation was explored using Principal Component Analysis (PCA), followed by formal statistical validation using MANOVA. Oil data analysis As seen in Fig. 6 , PCA of oil samples based on the selected ratiometric markers revealed discernible clustering, with the first two principal components explaining 76% and 15% of the total variance, respectively (91% cumulative variance). The grouping of oil types within the principal component space is evident in the corresponding score plot. Quantitative evaluation yielded a silhouette coefficient of 0.36 (Table 4 ), indicating weak-to-moderate cluster separation. In general silhouette coefficients range from − 1 to 0 to 1, with values close to zero reflecting overlapping clusters and values approaching unity indicating well-separated groups. In addition, the ratio of between-class to within-class scatter (B/W = 3.34) demonstrated that inter-class variability substantially exceeded intra-class variability in the PCA space. A permutation test on the B/W statistic (p_perm = 0.005) further confirmed that the observed clustering is unlikely to arise from random variation. Together, these complementary metrics indicate statistically meaningful separation between oil classes. Chips data analysis For chips, PCA revealed a reduced but still interpretable multivariate separation, with the first two principal components accounting for 57% and 20% of the variance, respectively (77% cumulative variance). Although cluster separation was less pronounced than in oils, quantitative measures still indicated non-random structure (silhouette = 0.12; B/W = 0.88; p_perm = 0.005). This reduced separation is expected to result from food-matrix interference, yet oil-dependent trends remain detectable in multivariate space which is valuable from screening purpose. While PCA and associated clustering metrics provide geometric and distributional measures of class separability, they do not constitute formal hypothesis tests. Therefore, multivariate analysis of variance (MANOVA) was employed to rigorously assess whether the observed group separation is statistically significant across all ratiometric variables simultaneously. Results obtained from the MANOVA analysis are presented in Table 4 . Table 4 Evaluation metrics by PCA and MANOVA PCA metrics Oils Chips Features Values Features Values Silhouette 0.361 Silhouette 0.116 Between_within_ratio 3.33 Between_within_ratio 0.87 p_perm 0.005 p_perm 0.005 MANOVA metrics Oils Chips Features Values F_value p Values F_value p Wilks' lambda 0.0008 984.83 < 0.001 0.0307 221.75 < 0.001 Pillai's trace 2.2241 207.28 < 0.001 1.5068 89.94 < 0.001 Hotelling-Lawley trace 184.68 7609.35 < 0.001 15.4828 573.41 < 0.001 Roy's greatest root 180.84 29929.62 < 0.001 14.4794 2155.02 < 0.001 For oils, Wilks’ lambda indicated strong multivariate separation among oil classes (Λ = 0.0008, F = 984.83, p < 0.001). Wilks’ lambda ranges from 0 to 1, with values close to unity indicating no group separation and values approaching zero reflecting near-complete discrimination. The observed value therefore indicates that only a negligible fraction of the total multivariate variance remains unexplained by oil type. Consistent results were obtained from Pillai’s trace (V = 2.2241, F = 207.28, p < 0.001), which ranges from 0 to minimum (p, k − 1), where (p = 5) is the number of variables and (k = 5) the number of groups, and quantifies the cumulative proportion of variance explained across multiple discriminant dimensions. In the present study, the observed Λ and V values correspond to more than half of the theoretical maximum, indicating strong multivariate separation. Similarly, the Hotelling–Lawley trace (T = 184.68, F = 7609.35, p < 0.001), which reflects the overall signal-to-noise ratio, and Roy’s greatest root (Θ = 180.84, F = 29929.62, p < 0.001), which measures separation along the strongest discriminant axis, both indicate highly pronounced group differences. For chips, despite reduced discriminatory power due to food-matrix interference, significant multivariate separation was likewise observed (Λ = 0.0307, V = 1.5068, T = 15.48, Θ = 14.48; all p < 0.0001). Although these values are smaller than those obtained for oils, they remain well within ranges associated with statistically meaningful discrimination. In particular, Wilks’ lambda remains close to zero and Pillai’s trace indicates substantial explained variance, demonstrating that oil-dependent spectral signatures are retained despite interference from non-lipid components. Taken together, the consistent significance across all four complementary statistics confirms robust and method-independent group separation. The close agreement between exploratory chemometric visualization (PCA) and formal multivariate inference (MANOVA) demonstrates that oil-type classification is driven by genuine multivariate oil-type-dependent compositional differences encoded in the ratiometric markers, establishing the statistical robustness and chemical relevance. Conclusion This study establishes a proof-of-concept Raman–chemometric framework for rapid authentication of edible oils in processed foods using chemically interpretable ratiometric markers. Key markers were systematically identified through a two-stage selection strategy combining Random Forest importance ranking and Kruskal–Wallis testing across all inter-peak combinations. Subsequent ANOVA and coefficient of variance (CV) analyses validated a final set of five robust markers exhibiting strong associations with saturated and unsaturated fatty acid content. PCA based chemometric analysis using these markers revealed pronounced clustering for oils (91% cumulative variance; silhouette = 0.36; B/W = 3.34) and moderate yet statistically significant separation for chips (77% cumulative variance; silhouette = 0.12; B/W = 0.88), with permutation testing confirming non-random structure (p = 0.005). Multivariate analysis of variance further validated group separation for both oils (Wilks’ Λ = 0.0008, p < 0.001) and chips (Wilks’ Λ = 0.0307, p < 0.0001), with consistent significance across complementary test statistics, demonstrating that the observed discrimination reflects genuine oil-type-dependent multivariate compositional differences rather than sampling variability. In parallel, this study introduces a novel tissue paper–based minimal sampling strategy tailored for complex food matrices like fried chips, providing a practical and environmentally sustainable alternative to conventional solvent-intensive extraction protocols. This approach facilitates rapid, low-cost sample preparation while preserving analytical sensitivity in multi-component systems. Overall, the proof-of-concept framework provides a foundation for automated and AI-assisted oil authentication workflow in processed foods. While the present study was conducted under controlled laboratory conditions using a limited number of oil types and a single food matrix, the robustness and chemical interpretability of the proposed framework provide a strong basis for extension to diverse industrial processing conditions, oil blends, varied adulteration scenarios, and commercial products. Future work will focus on expanding the methodology to a broader range of oil varieties and diverse food matrices. In addition, integration with advanced machine learning models will be explored to further enhance classification accuracy and enable scalable, real-time food authenticity monitoring and quality control. Declarations Acknowledgements The authors would like to express their sincere gratitude to the Founder Chancellor and the Institute Management of Sri Sathya Sai Institute of Higher Learning, Anantapur, Andhra Pradesh, India, for their support in facilitating this research. The authors also acknowledge the Central Research Instrumentation Facility (CRIF) for providing the essential resources and facilities necessary for the instrumental analysis. Additionally, the authors extend their thanks to the Department of Food and Nutritional Sciences for their valuable resources and support throughout the course of this work. Author contribution J.G. conceptualization, supervision, formal analysis, review & editing, approval of final version of manuscript. A.S. conceptualization, methodology, investigation, software, data curation, formal analysis, visualization, writing – original draft, approval of final version of manuscript. C.S.N. software, data curation, formal analysis, visualization, review & editing, approval of final version of manuscript. S.M.V. supervision, review & editing, approval of final version of manuscript. D.L.N.K. software, formal analysis, review & editing, approval of final version of manuscript. B.P.R. data curation, formal analysis, approval of final version of manuscript. Data Availability The datasets and analysis scripts generated and used in the present study are available from the corresponding author upon reasonable request and will be shared on a case-by-case basis, subject to ongoing related research activities and institutional policies. Competing interests The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Funding The authors did not receive support from any organization for the submitted work. References Alkarkhi, A. F. M., Alqaraghuli, W. A. A. (2019). Comparing several population means. Easy Statistics for Food Science with R. Academic Press, pp 81-105. https://doi.org/10.1016/B978-0-12-814262-2.00006-6 Aparicio R, Aparicio-Ruı́z R (2000) Authentication of vegetable oils by chromatographic techniques. Journal of Chromatography A 881:93–104. https://doi.org/10.1016/S0021-9673(00)00355-1 Aykas DP, Rodriguez-Saona LE (2016) Assessing potato chip oil quality using a portable infrared spectrometer combined with pattern recognition analysis. Anal Methods 8:731–741. https://doi.org/10.1039/C5AY02387D Baeten V, Hourant P, Morales MT, Aparicio R (1998) Oil and Fat Classification by FT-Raman Spectroscopy. J Agric Food Chem 46:2638–2646. https://doi.org/10.1021/jf9707851 Bansal G, Zhou W, Barlow PJ, et al (2010) Performance of palm olein in repeated deep frying and controlled heating processes. Food Chemistry 121:338–347. https://doi.org/10.1016/j.foodchem.2009.12.034 Barton SJ, Ward TE, Hennelly BM (2018) Algorithm for optimal denoising of Raman spectra. Anal Methods 10:3759–3769. https://doi.org/10.1039/C8AY01089G Begum A, Jain BP (2024) Adulteration in edible oil (mustard oil) and ghee; detection and their effects on human health. Int J Biochem Mol Biol 15:141–148. https://doi.org/10.62347/VVCI2092 Comparing Several Population Means. https://www.sciencedirect.com/science/chapter/monograph/abs/pii/B9780128142622000066. Accessed 4 Feb 2026 Castro RC, Ribeiro DSM, Santos JLM, Páscoa RNMJ (2022) The use of in-situ Raman spectroscopy to monitor at real time the quality of different types of edible oils under frying conditions. Food Control 136:108879. https://doi.org/10.1016/j.foodcont.2022.108879 Cercaci L, Rodriguezestrada M, Lercker G (2003) Solid-phase extraction–thin-layer chromatography–gas chromatography method for the detection of hazelnut oil in olive oils by determination of esterified sterols. Journal of Chromatography A 985:211–220. https://doi.org/10.1016/S0021-9673(02)01397-3 Davari M, Bahreini M, Sabzevari Z (2024) Developing a non-destructive method for the detection of egg quality and freshness using micro-Raman spectroscopy. Applied Food Research 4:100453. https://doi.org/10.1016/j.afres.2024.100453 Desk, T. L. (2024, June 6). How to identify fake cooking oil? (FSSAI recommended tips). The Times of India. https://timesofindia.indiatimes.com/life-style/health-fitness/health-news/how-to-identify-fake-cooking-oil-fssai-recommended-tips/articleshow/110694766.cms Dorni C, Sharma P, Saikia G, Longvah T (2018) Fatty acid profile of edible oils and fats consumed in India. Food Chemistry 238:9–15. https://doi.org/10.1016/j.foodchem.2017.05.072 Elumalai A, Natarajan V (2025) Advancements in analytical technologies for ensuring food quality and authentication: A comprehensive review. Journal of Food Composition and Analysis 139:107075. https://doi.org/10.1016/j.jfca.2024.107075 Fatima S, Kumar V, Bhadauria G, Verma H (2023) Quality indicators based rapid test kits for detection of frying oil quality: A review. Food Chemistry Advances 2:100305. https://doi.org/10.1016/j.focha.2023.100305 Forooghi E (2023) Applications of Raman spectroscopy in food safety and quality control. HHHM 4:2. https://doi.org/10.30502/jhhhm.2024.430203.1082 Foster R, Williamson CS, Lunn J (2009) BRIEFING PAPER: Culinary oils and their health effects. Nutrition Bulletin 34:4–47. https://doi.org/10.1111/j.1467-3010.2008.01738.x Gómez-Caravaca AM, Maggio RM, Cerretani L (2016) Chemometric applications to assess quality and critical parameters of virgin and extra-virgin olive oil. A review. Analytica Chimica Acta 913:1–21. https://doi.org/10.1016/j.aca.2016.01.025 Gong W, Shi R, Chen M, et al (2019) Quantification and monitoring the heat-induced formation of trans fatty acids in edible oils by Raman Spectroscopy. Food Measure 13:2203–2210. https://doi.org/10.1007/s11694-019-00140-5 Gu H, Huang X, Sun Y, et al (2021) Intelligent evaluation of total polar compounds (TPC) content of frying oil based on fluorescence spectroscopy and low-field NMR. Food Chemistry 342:128242. https://doi.org/10.1016/j.foodchem.2020.128242 Haji A, Desalegn K, Hassen H (2023) Selected food items adulteration, their impacts on public health, and detection methods: A review. Food Sci Nutr 11:7534–7545. https://doi.org/10.1002/fsn3.3732 Hu R, He T, Zhang Z, et al (2019) Safety analysis of edible oil products via Raman spectroscopy. Talanta 191:324–332. https://doi.org/10.1016/j.talanta.2018.08.074 Huang F, Li Y, Guo H, et al (2016) Identification of waste cooking oil and vegetable oil via Raman spectroscopy. J Raman Spectroscopy 47:860–864. https://doi.org/10.1002/jrs.4895 Huang X, Li H, Ruan Y, et al (2022) An integrated approach utilizing raman spectroscopy and chemometrics for authentication and detection of adulteration of agarwood essential oils. Front Chem 10:1036082. https://doi.org/10.3389/fchem.2022.1036082 Inteaz A (1989) Food Quality Assurance: Elements of Raw Materials Control. Canadian Institute of Food Science and Technology Journal 22:113–115. https://doi.org/10.1016/S0315-5463(89)70337-0 Jamieson LE, Li A, Faulds K, Graham D (2018a) Ratiometric analysis using Raman spectroscopy as a powerful predictor of structural properties of fatty acids. R Soc Open Sci 5:181483. https://doi.org/10.1098/rsos.181483 Jamieson LE, Li A, Faulds K, Graham D (2018b) Ratiometric analysis using Raman spectroscopy as a powerful predictor of structural properties of fatty acids. R Soc Open Sci 5:181483. https://doi.org/10.1098/rsos.181483 Johnson DS, Saikia DN (2009) Fatty acids profile of Edible Oils and Fats in India Jolliffe IT, Cadima J (2016) Principal component analysis: a review and recent developments. Phil Trans R Soc A 374:20150202. https://doi.org/10.1098/rsta.2015.0202 Kaimal A, Dhingra M, Singhal R (2021) Monitoring of oil quality from commercial fried foods‐A case study from India. Journal of Food Processing and Preservation 46:. https://doi.org/10.1111/jfpp.16138 Kaur A, Scarborough P, Hieke S, et al (2016) The nutritional quality of foods carrying health-related claims in Germany, The Netherlands, Spain, Slovenia and the United Kingdom. European Journal of Clinical Nutrition 70:114. https://doi.org/10.1038/ejcn.2016.114 Khursheed M (2024) Chromatographic Techniques for the Detection and Identification of Olive Oil Adulteration. ReiDoCrea: Revista electrónica de investigación Docencia Creativa. https://doi.org/10.30827/Digibug.86578 Kim N, Yu KS, Kim J, et al (2018) Chemical characteristics of potato chips fried in repeatedly used oils. Food Measure 12:1863–1871. https://doi.org/10.1007/s11694-018-9800-x Korepanov VI (2020) Asymmetric least-squares baseline algorithm with peak screening for automatic processing of the Raman spectra. Journal of Raman Spectroscopy 51:2061–2065. https://doi.org/10.1002/jrs.5952 Kwofie F, Lavine BK, Ottaway J, Booksh K (2020) Differentiation of Edible Oils by Type Using Raman Spectroscopy and Pattern Recognition Methods. Appl Spectrosc 74:645–654. https://doi.org/10.1177/0003702819888220 Liu H, Chen Y, Shi C, et al (2020) FT-IR and Raman spectroscopy data fusion with chemometrics for simultaneous determination of chemical quality indices of edible oils during thermal oxidation. LWT 119:108906. https://doi.org/10.1016/j.lwt.2019.108906 Mendes TO, Da Rocha RA, Porto BLS, et al (2015) Quantification of Extra-virgin Olive Oil Adulteration with Soybean Oil: a Comparative Study of NIR, MIR, and Raman Spectroscopy Associated with Chemometric Approaches. Food Anal Methods 8:2339–2346. https://doi.org/10.1007/s12161-015-0121-y Menze BH, Kelm BM, Masuch R, et al (2009) A comparison of random forest and its Gini importance with standard chemometric methods for the feature selection and classification of spectral data. BMC Bioinformatics 10:213. https://doi.org/10.1186/1471-2105-10-213 Moore JC, Spink J, Lipp M (2012) Development and Application of a Database of Food Ingredient Fraud and Economically Motivated Adulteration from 1980 to 2010. Journal of Food Science 77:R118–R126. https://doi.org/10.1111/j.1750-3841.2012.02657.x Navratilova K, Hurkova K, Hrbek V, Uttl L (2022) Metabolic fingerprinting strategy: Investigation of markers for the detection of extra virgin olive oil adulteration with soft-deodorized olive oils. Food Control 134: 108649. https://doi.org/10.1016/j.foodcont.2021.108649 Pereira FJ, López R, Ferrer N, et al (2021) A comparative appraisal of Raman band ratioing and chemometric analysis for classification of ancient papyri. Journal of Cultural Heritage 52:55–64. https://doi.org/10.1016/j.culher.2021.09.003 Program HF (2024) Economically Motivated Adulteration (Food Fraud) Research Publications. FDA. https://www.fda.gov/food/compliance-enforcement-food/economically-motivated-adulteration food-fraud Rana R, Kumar N, Pandit A, et al (2014) Effect of various socio-economic factors on the consumption of processed potato products in Punjab. Indian Journal of Agricultural Marketing 28:24–35. https://www.researchgate.net/profile/Rajesh-Rana 3/publication/264889204_Effect_of_various_socio economic_factors_on_the_consumption_of_processed_potato_products_in_Punjab/links/5771046 808ae6219474a3263/Effect-of-various-socio-economic-factors-on-the-consumption-of-processed potato-products-in-Punjab.pdf Rifna EJ, Pandiselvam R, Kothakota A, et al (2022) Advanced process analytical tools for identification of adulterants in edible oils – A review. Food Chemistry 369:130898. https://doi.org/10.1016/j.foodchem.2021.130898 Sharma R (2022). Food Safety and Standards Authority of India Press release, India. Retrieved from https.://www.fssai.gov.in/upload/press_release/2022/08/62fcb62aa2949Press_Release_Edible_oil_ 17_08_2022.pdf Shehata M, Dodd S, Mosca S, et al (2024) Application of Spatial Offset Raman Spectroscopy (SORS) and Machine Learning for Sugar Syrup Adulteration Detection in UK Honey. Foods 13:2425. https://doi.org/10.3390/foods13152425 Su R, Xu X, Wang X, et al (2011) Determination of organophosphorus pesticides in peanut oil by dispersive solid phase extraction gas chromatography–mass spectrometry. Journal of Chromatography B 879:3423–3428. https://doi.org/10.1016/j.jchromb.2011.09.016 Tan C-H, Kong I, Irfan U, et al (2021) Edible Oils Adulteration: A Review on Regulatory Compliance and Its Detection Technologies. Journal of Oleo Science 70:1343–1356. https://doi.org/10.5650/jos.ess21109 Ün İ, OK S (2018) Analysis of olive oil for authentication and shelf life determination. J Food Sci Technol 55:2476–2487. https://doi.org/10.1007/s13197-018-3165-3 Vaskova H, Buckova M (2018) Multivariate study of Raman spectral data of edible oils. 14: 02022. https://doi.org/10.1051/matecconf/201712502022 Wang J, Qian J, Xu M, et al (2025) Adulteration detection of multi-species vegetable oils in camellia oil using Raman spectroscopy: Comparison of chemometrics and deep learning methods. Food Chemistry 463:141314. https://doi.org/10.1016/j.foodchem.2024.141314 Wu Y, Liu J, Xu R, et al (2024) Recent advances in ratiometric surface-enhanced Raman spectroscopy sensing strategies. Microchemical Journal 199:110127. https://doi.org/10.1016/j.microc.2024.110127 Yadav S (2018) Edible oil adulterations: Current issues, detection techniques, and health hazards. Int J Chem Stud 6:1393–1397 Yang T, Zhao B, He L (2019) Raman instruments for food quality evaluation. In: Evaluation Technologies for Food Quality. Elsevier, pp 119–143. https://www.chemijournal.com/archives/2018/vol6issue2/PartT/6-2-102-397.pdf Yao S, Aykas DP, Rodriguez-Saona L (2020) Rapid Authentication of Potato Chip Oil by Vibrational Spectroscopy Combined with Pattern Recognition Analysis. Foods 10:42. https://doi.org/10.3390/foods10010042 Zhu P, Yang Q, Zhao H (2022) Identification of peanut oil origins based on Raman spectroscopy combined with multivariate data analysis methods. Journal of Integrative Agriculture 21:2777–2785. https://doi.org/10.1016/j.jia.2022.07.026 Additional Declarations No competing interests reported. Supplementary Files Graphicalabstract.docx Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 12 Mar, 2026 Reviews received at journal 04 Mar, 2026 Reviews received at journal 25 Feb, 2026 Reviewers agreed at journal 20 Feb, 2026 Reviewers agreed at journal 20 Feb, 2026 Reviewers invited by journal 19 Feb, 2026 Editor assigned by journal 15 Feb, 2026 Submission checks completed at journal 15 Feb, 2026 First submitted to journal 13 Feb, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8872116","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":594309272,"identity":"342ae880-f024-4588-937c-e3213349b72c","order_by":0,"name":"Amrita Shaw","email":"","orcid":"","institution":"Sri Sathya Sai Institute of Higher Learning","correspondingAuthor":false,"prefix":"","firstName":"Amrita","middleName":"","lastName":"Shaw","suffix":""},{"id":594309273,"identity":"5b7a4655-0c28-4ec8-9543-d93f516c86a1","order_by":1,"name":"Chandrasekar SUBRAMANI NARAYANA","email":"","orcid":"","institution":"Aix-Marseille University","correspondingAuthor":false,"prefix":"","firstName":"Chandrasekar","middleName":"SUBRAMANI","lastName":"NARAYANA","suffix":""},{"id":594309274,"identity":"b1a31a5e-725b-40d7-a707-53d7b3816777","order_by":2,"name":"Sai Muthukumar V","email":"","orcid":"","institution":"Sri Sathya Sai Institute of Higher Learning","correspondingAuthor":false,"prefix":"","firstName":"Sai","middleName":"Muthukumar","lastName":"V","suffix":""},{"id":594309275,"identity":"ce51df75-6f5a-4acf-9e23-27d08a43ec13","order_by":3,"name":"Deepak L N Kallepalli","email":"","orcid":"","institution":"CogniEvolve AI Inc","correspondingAuthor":false,"prefix":"","firstName":"Deepak","middleName":"L N","lastName":"Kallepalli","suffix":""},{"id":594309276,"identity":"223123c7-3885-4eb4-a881-c35ace83df87","order_by":4,"name":"Bhanu Prakash Rachaiah","email":"","orcid":"","institution":"Science4u Analytics And Research Solutions Private Limited","correspondingAuthor":false,"prefix":"","firstName":"Bhanu","middleName":"Prakash","lastName":"Rachaiah","suffix":""},{"id":594309277,"identity":"8d59fb8b-7aa1-4500-a637-0944f8fc49d1","order_by":5,"name":"Jhinuk Gupta","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABA0lEQVRIiWNgGAWjYBACxhmMDQwMByDsB0CCx4AdRIIRIS1sDMwGYC3MBLQwSIAIiBY2MBumBSdgnt3c9pjnTG3idvnmZ5U//tTJmDMDXfi2jUHGHJfD5hxsN+a5cTxxZxub2W3etsM8ls0MzIZz2xh4LBtw+SWxTZrnw7HEDccYzG4zNhzgMTjMwCbNC9RicICgFvZvhUCHgbSw/yas5UYNUAuPGQMPGzPYFmZCWiTnnDlgvLMtp1ga5BeDw4zNknPOSeDUYjgj/ZnEm2N1stuZj2/8CHSYvcHx5oMf3pTZ2OPUAgmWwwwGSDaDxCSwqwcCeQhVh6xlFIyCUTAKRgEqAACPhVtekJe0yAAAAABJRU5ErkJggg==","orcid":"","institution":"Sri Sathya Sai Institute of Higher Learning","correspondingAuthor":true,"prefix":"","firstName":"Jhinuk","middleName":"","lastName":"Gupta","suffix":""}],"badges":[],"createdAt":"2026-02-13 12:54:51","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8872116/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8872116/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":103506053,"identity":"9c3967fd-7408-4f2a-a776-b828228fe5f8","added_by":"auto","created_at":"2026-02-26 13:33:56","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":389094,"visible":true,"origin":"","legend":"\u003cp\u003eSteps followed in sample preparation (with Sunflower Oil as an example).300 mL of Sunflower oil was heated to frying temperature (195˚ C). Thin potato slices were added to the oil, and fried until golden brown. The fried chips were cooled to room temperature and collected in an airtightpouch, marked with the sample codes. After cooling the oil in the pan, aliquots of 5 mL (used oil) were collected in glass vials and were sequentially labelled in both oil and chips category (additional suffix ‘C’ for chips category).\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-8872116/v1/d83d35a630a0e0ce364aaace.png"},{"id":103506843,"identity":"10b76f6c-0543-40ca-82ae-86a4e2ad53e4","added_by":"auto","created_at":"2026-02-26 13:39:41","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":214592,"visible":true,"origin":"","legend":"\u003cp\u003eSchematic illustration of the developed tissue paper-based extraction method for the food matrix.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-8872116/v1/4e203f2fc9e494f02a889c66.png"},{"id":103324200,"identity":"9ef6e50b-5134-47eb-bf46-e6c9e7784d79","added_by":"auto","created_at":"2026-02-24 12:33:20","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":107827,"visible":true,"origin":"","legend":"\u003cp\u003eSchematic flow for spectral data processing used in the proposed framework\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-8872116/v1/2df91927bb2e333686e815ac.png"},{"id":103507172,"identity":"9004b529-7f1d-4260-9571-064fb3a3c09d","added_by":"auto","created_at":"2026-02-26 13:40:39","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":362887,"visible":true,"origin":"","legend":"\u003cp\u003eSchematic workflow illustrating the development and validation of the ratiometric markers in the current framework.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-8872116/v1/9d6dcb1abd1e48139a6368e9.png"},{"id":103324201,"identity":"64b885b4-ef18-47dd-ae5b-dcdf68053bed","added_by":"auto","created_at":"2026-02-24 12:33:20","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":277952,"visible":true,"origin":"","legend":"\u003cp\u003eBaseline corrected, smoothened and normalized\u003cstrong\u003e \u003c/strong\u003eRaman spectra collected from oils and chips\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-8872116/v1/fb6a6d1694ba037a15f24600.png"},{"id":103324203,"identity":"128acdbd-3ac1-4ff2-ae0b-03ee121c4ee6","added_by":"auto","created_at":"2026-02-24 12:33:20","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":363735,"visible":true,"origin":"","legend":"\u003cp\u003ePCA plot from oil and chips samples.\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-8872116/v1/9a6c5242081092f7f163df48.png"},{"id":103511387,"identity":"24520d70-9abb-4690-8841-84edccbbde19","added_by":"auto","created_at":"2026-02-26 14:09:33","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2843951,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8872116/v1/f5b80a36-298a-44c2-af58-b9ec22f872f1.pdf"},{"id":103324204,"identity":"d0aec580-657b-4713-b641-85970c47ca95","added_by":"auto","created_at":"2026-02-24 12:33:20","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":270227,"visible":true,"origin":"","legend":"","description":"","filename":"Graphicalabstract.docx","url":"https://assets-eu.researchsquare.com/files/rs-8872116/v1/470d6a32149f35100ed1cf28.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Raman-Chemometric Framework for Rapid Authentication of Edible Oils in Processed Foods Using a Validated One Step Sampling Technique","fulltext":[{"header":"Introduction","content":"\u003cp\u003eIdentifying the type of oil used in processed foods is crucial due to its significant impact on consumer health (Foster et al. \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; Moore et al. \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). A common form of adulteration involves substituting listed oils with cheaper analogues, locally available alternatives, or repeatedly heated oils, compromising both food quality and safety (Rifna et al. \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Haji et al. \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Some of these substitutes, such as argemone oil, contain toxins and are banned globally by regulatory agencies (Tan et al. \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Begum and Jain \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Even legally permitted oils like palm oil and vanaspati pose health risks and are unsuitable for individuals with cardiovascular diseases and various health issues, while soybean and groundnut oils can trigger allergic reactions (Johnson and Saikia \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; Dorni et al. \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Although oil adulteration is a global issue, it becomes a formidable challenge in developing countries like India (Rana et al. \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Regulatory enforcement mainly targets the organized food sector, leaving much of the unregulated market vulnerable. As a result, incidents of such adulteration have increased significantly in recent years (Program \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Moreover, recent investigations suggest that economically motivated adulteration is no longer confined to unregistered businesses. A nationwide audit by the Food Safety and Standards Authority of India (FSSAI) in 2022, seized thousands of liters of edible oils for suspected mislabeling, fake branding and exceeding levels of trans-fatty acids, highlighting the urgent need for stricter monitoring and legal intervention (Sharma R, \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Desk, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Typically, Chromatographic methods like High-Performance Liquid Chromatography (HPLC) (Cercaci et al. \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2003\u003c/span\u003e; Navratilova et al. \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Khursheed \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) and Gas Chromatography (GC) (Aparicio and Aparicio-Ruı́z \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2000\u003c/span\u003e; Su et al. \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2011\u003c/span\u003e; \u0026Uuml;n and OK \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2018\u003c/span\u003e), along with Nuclear Magnetic Resonance (NMR) (\u0026Uuml;n and OK \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Gu et al. \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), are used for oil authentication. While these methods provide high accuracy, their complex sample processing protocols, lengthy analysis procedures and high operational costs make them unsuitable for rapid, real time monitoring (Elumalai and Natarajan \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Thus, a high-throughput screening technique with minimal sample processing is essential and a desirable alternative.\u003c/p\u003e \u003cp\u003eVibrational spectroscopic techniques have emerged as a superior alternative to chromatographic methods, overcoming their limitations and enabling real-time monitoring of edible oils (Yadav \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Hu et al. \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Yao et al. \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Forooghi \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Among these, Raman spectroscopy stands out for its ability to provide detailed molecular fingerprints \u003cem\u003evia\u003c/em\u003e Raman scattering (Mendes et al. \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Yang et al. \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Forooghi \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). It offers a promising alternative because it directly probes molecular vibrations associated with lipid structure, including saturation, unsaturation and ester functionality (Baeten et al. \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e1998\u003c/span\u003e; Huang et al. \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Kwofie et al. \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Wang et al. \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Characteristic Raman bands at critical wavenumbers like 1650\u0026ndash;1660 cm⁻\u0026sup1; (C\u0026thinsp;=\u0026thinsp;C), 1740\u0026ndash;1750 cm⁻\u0026sup1; (ester carbonyl), and 1430\u0026ndash;1440 cm⁻\u0026sup1; (CH₂ scissoring) have been widely used to differentiate oil types and monitor oxidation (Huang et al. \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Wang et al. \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Attempts have been made to probe the quality of waste cooking oils (Huang et al. \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2016\u003c/span\u003e) and the quantification of trans-fat in edible oils (Gong et al. \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) using Raman spectroscopy. Therefore, the current trends in food quality monitoring clearly indicate the adaption of molecular spectroscopic techniques like Raman Spectrometry for the development of rapid quality monitoring tools.\u003c/p\u003e \u003cp\u003eRatiometric analysis in spectroscopy involves calculating the ratio of intensities between two characteristic peaks rather than using absolute intensity values, that enhances measurement sensitivity and selectivity (Wu et al. \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Ratiometric interpretation of Raman bands is increasingly common due to its robustness against intensity fluctuations, laser defocus, and sample variability (Jamieson et al. 2018). However, the literature lacks ratiometric markers that are systematically derived, reproducible, statistically validated, and correlated with chemical composition. Stringent selection of these ratiometric markers helps to reduce the high dimensionality of spectral data and improve statistical performance (Jamieson et al. 2018). Unfortunately, most existing studies rely only on visual or empirical selection of ratios without rigorous statistical validation. Therefore, it is important to develop a systematic framework for identifying and validating ratiometric markers through quantitative correlation analysis and multivariate statistical testing.\u003c/p\u003e \u003cp\u003eThe application of multivariate analysis to spectral data highlights the growing importance of data analysis in food quality monitoring (Vaskova and Buckova \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Chemometric analysis combined with spectrometry has been widely recognized for its effectiveness, particularly in oil quality assessment (G\u0026oacute;mez-Caravaca et al. \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Integrating vibrational spectroscopic techniques with chemometric methods is expected to further enhance the evaluation of edible oil quality (Liu et al. \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Rigorous application of chemometrics for oil authentication in complex food matrices is gap in the scientific domain which demands more efforts.\u003c/p\u003e \u003cp\u003eQuality monitoring in processed food is more challenging than quality monitoring in ingredients due to matrix interference. One such example is interference from starch and seasonings while screening for the oil in potato chips (food matrix). Most of the food quality monitoring studies deal with ingredients testing rather than testing the final food product (Inteaz \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e1989\u003c/span\u003e) as testing of such end products involves elaborate sampling, extraction, isolation, and derivatization. Studies have shown that crucial components, such as fats and oils, which are extensively used in food preparation still need to be isolated from matrices (Bansal et al. \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2010\u003c/span\u003e) \u003cem\u003evia\u003c/em\u003e a long, tedious process before impending analysis (Kim et al. \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Yao et al. \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Kaimal et al. \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). As per example, the traditional method for oil analysis in fried food matrix involves mechanical pressing to extract oil followed by debris removal, filtration and additional purification steps (Aykas and Rodriguez-Saona \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). This extended processing time underscores the need for developing a rapid, high throughput, inexpensive and efficient sampling protocol.\u003c/p\u003e \u003cp\u003eAs seen above, most of the reported studies have explored oil type analysis using Raman-PCA (Castro et al. \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) involving tedious traditional multi-step processes. Little efforts have been made on developing techniques with minimal sample processing, directly from the food matrix, which can actually make the oil quality monitoring highly reliable and reproducible. From a food safety perspective, a framework capable of analyzing oil types from food matrices is inevitably crucial.\u003c/p\u003e \u003cp\u003eIn this project, we have developed Raman-PCA frameworks to identify oil types in processed foods (Potato chips as the food matrix) with minimal sample processing. A novel, non-invasive tissue paper-based oil sampling technique for food matrix was developed and validated in the study. Potato chips fried in five different edible oils, including sunflower oil, vanaspati oil, palm oil, groundnut oil, and soybean oil, were explored to design the framework. While these oils are not always classified as adulterants, they can pose health risks to particular individuals with soy or peanut allergies or cardiovascular conditions when added undeclared. For developing the framework, spectral analysis identified fatty acid-associated Raman bands from which inter-peak intensity ratios were systematically constructed. A two-stage selection strategy combining Random Forest marker importance ranking and Kruskal-Wallis statistical testing identified the most discriminatory intensity ratios, which were further validated through one-way ANOVA and coefficient of variance (CV) analysis for marker reproducibility and robustness. The selected ratiometric markers were subjected to multivariate analysis, with principal component analysis (PCA) performed to assess class separability, followed by formal statistical validation using MANOVA to confirm the robustness of multivariate separation beyond visual inspection.\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eProcurement and preparation of samples\u003c/h2\u003e \u003cp\u003eFive commercially available packaged edible oils (Sunflower-SO, Soybean-SOYO, Groundnut-GNO, Palm-PO, and Vanaspati-VO) with FSSAI certification were procured from retail outlets in Anantapur district, Andhra Pradesh, India. Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e depicts the steps employed for SO, which was systematically replicated across all oil types. The samples, along with their codes, are provided in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. To mimic actual practices followed by small scale potato chips manufacturing units, thin slices of potatoes were fried in each oil over 9 successive batches, with systematic collection of both fried chips and oil samples.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eList of samples prepared for analysis\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"12\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c11\" colnum=\"11\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c12\" colnum=\"12\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"10\" nameend=\"c12\" namest=\"c3\"\u003e \u003cp\u003eSample Codes\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHeating Stages\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"3\" nameend=\"c4\" namest=\"c2\"\u003e \u003cp\u003eSunflower Oil (SO)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003ePalm Oil (PO)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003eVanaspati (VO)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003eSoya Bean Oil (SOYO)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c12\" namest=\"c11\"\u003e \u003cp\u003eGroundnut Oil (GNO)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eOil\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eFood matrix (Chips)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eOil\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eFood matrix (Chips)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eOil\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eFood matrix (Chips)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eOil\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003eFood matrix (Chips)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c11\"\u003e \u003cp\u003eOil\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c12\"\u003e \u003cp\u003eFood matrix (Chips)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e\u003cem\u003eSO 0\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003e-\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003ePO 0\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cem\u003e-\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cem\u003eVO 0\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u003cem\u003e-\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cem\u003eSOYO 0\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u003cem\u003e-\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u003cem\u003eGNO 0\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e\u003cem\u003e-\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e\u003cem\u003eSO 1\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eSO C 1\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003ePO 1\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cem\u003ePO C 1\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cem\u003eVO 1\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u003cem\u003eVO C 1\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cem\u003eSOYO 1\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u003cem\u003eSOYO C 1\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u003cem\u003eGNO 1\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e\u003cem\u003eGNO C 1\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003e2\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e\u003cem\u003eSO 2\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eSO C 2\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003ePO 2\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cem\u003ePO C 2\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cem\u003eVO 2\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u003cem\u003eVO C 2\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cem\u003eSOYO 2\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u003cem\u003eSOYO C 2\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u003cem\u003eGNO 2\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e\u003cem\u003eGNO C 2\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003e3\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e\u003cem\u003eSO 3\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eSO C 3\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003ePO 3\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cem\u003ePO C 3\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cem\u003eVO 3\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u003cem\u003eVO C 3\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cem\u003eSOYO 3\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u003cem\u003eSOYO C 3\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u003cem\u003eGNO 3\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e\u003cem\u003eGNO C 3\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003e4\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e\u003cem\u003eSO 4\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eSO C 4\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003ePO 4\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cem\u003ePO C 4\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cem\u003eVO 4\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u003cem\u003eVO C 4\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cem\u003eSOYO 4\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u003cem\u003eSOYO C 4\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u003cem\u003eGNO 4\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e\u003cem\u003eGNO C 4\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003e5\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e\u003cem\u003eSO 5\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eSO C 5\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003ePO 5\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cem\u003ePO C 5\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cem\u003eVO 5\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u003cem\u003eVO C 5\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cem\u003eSOYO 5\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u003cem\u003eSOYO C 5\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u003cem\u003eGNO 5\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e\u003cem\u003eGNO C 5\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003e6\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e\u003cem\u003eSO 6\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eSO C 6\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003ePO 6\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cem\u003ePO C 6\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cem\u003eVO 6\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u003cem\u003eVO C 6\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cem\u003eSOYO 6\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u003cem\u003eSOYO C 6\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u003cem\u003eGNO 6\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e\u003cem\u003eGNO C 6\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003e7\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e\u003cem\u003eSO 7\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eSO C 7\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003ePO 7\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cem\u003ePO C 7\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cem\u003eVO 7\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u003cem\u003eVO C 7\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cem\u003eSOYO 7\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u003cem\u003eSOYO C 7\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u003cem\u003eGNO 7\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e\u003cem\u003eGNO C 7\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003e8\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e\u003cem\u003eSO 8\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eSO C 8\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003ePO 8\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cem\u003ePO C 8\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cem\u003eVO 8\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u003cem\u003eVO C 8\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cem\u003eSOYO 8\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u003cem\u003eSOYO C 8\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u003cem\u003eGNO 8\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e\u003cem\u003eGNO C 8\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003e9\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e\u003cem\u003eSO 9\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eSO C 9\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003ePO 9\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cem\u003ePO C 9\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cem\u003eVO 9\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u003cem\u003eVO C 9\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cem\u003eSOYO 9\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u003cem\u003eSOYO C 9\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u003cem\u003eGNO 9\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e\u003cem\u003eGNO C 9\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eTissue-paper based oil sampling for food matrix (potato chips)\u003c/h3\u003e\n\u003cp\u003eA novel, simplistic but promising tissue paper\u0026ndash;based sampling technique tailored for food matrices was developed and evaluated in this study, with potato chips used as a representative model system as a food matrix. The proposed extraction approach is intended as a rapid and facile screening tool for oil authentication. Single-ply, laboratory-grade, non-fragrant tissue paper (Kimwipes\u0026reg;, Kimberly-Clark) was cut into uniform sections and placed between two fried potato chips with consistent pressure to facilitate oil absorption through capillary action. This method enabled efficient transfer of surface oil into the tissue paper \u003cem\u003evia\u003c/em\u003e blotting. These tissue papers which assimilated oil from the fried chips were subsequently analyzed directly by Raman spectroscopy without any further processing. A pictorial depiction of the developed extraction technique is presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003eData acquisition\u003c/h3\u003e\n\u003cp\u003eRaman spectroscopic measurements of all oil and chip samples were performed using a dispersive Raman microscope system (DXR, Thermo Fisher Scientific) equipped with a 785 nm excitation laser. After preliminary spectral acquisition survey, the instrumental parameters were optimized as follows: laser power, 20 mW; exposure time, 30 s; number of exposures, 2; confocal aperture, 25 \u0026micro;m slit; diffraction grating, 400 lines mm⁻\u0026sup1;; spectral resolution, 4 cm⁻\u0026sup1;. Raman spectra were acquired over the typical molecular finger printing range of 200\u0026ndash;3000 cm⁻\u0026sup1;. To ensure statistical reliability and account for sample heterogeneity, 20 different spectra were collected by focusing on different locations on each sample. This spatially resolved acquisition strategy enhances measurement robustness and reduces local sampling bias. The resulting spectra were treated as technical replicates representative of each experimental condition, and all subsequent analyses were performed accordingly under controlled laboratory settings. The complete dataset comprised 1000 Raman spectra from oil samples (200 spectra per oil type across 10 heating stages, including pristine) and 900 spectra from chip samples (180 spectra per oil type across 9 frying stages, starting from 1st frying as chips cannot exist in pristine form), totaling 1900 spectra across five oil types.\u003c/p\u003e\n\u003ch3\u003eData preprocessing\u003c/h3\u003e\n\u003cp\u003e \u003c/p\u003e \u003cp\u003eThe spectral dataset was subjected to a systematic preprocessing pipeline comprising three sequential steps (Fig.\u0026nbsp;3). (i) Baseline correction was performed using the asymmetric least squares (ALS) (Korepanov \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) algorithm with optimized parameters: asymmetry parameter (p)\u0026thinsp;=\u0026thinsp;0.01, smoothing parameter (λ)\u0026thinsp;=\u0026thinsp;10⁵, and 10 iterations. (ii) Spectral smoothing was applied using the Savitzky\u0026ndash;Golay filter (Barton et al. \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) with an 11-point window and third-order polynomial fitting to reduce noise while preserving spectral features. (iii) Finally, the entire dataset was normalized to the 2720 cm⁻\u0026sup1; band, which corresponds to C\u0026ndash;H stretching vibrations found commonly in all fatty acid chains across all samples. Specifically, 2720 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e was chosen as the wavenumber for normalization as it will not hamper the relative intensities of other functionally important peaks (Baeten et al. \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e1998\u003c/span\u003e; Kwofie et al. \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cb\u003eFigure\u0026nbsp;3\u003c/b\u003e Schematic flow for spectral data processing used in the proposed framework\u003c/p\u003e\n\u003ch3\u003eConstruction of ratiometric markers\u003c/h3\u003e\n\u003cp\u003eThe key component of identifying and differentiating various oil in the proposed framework is attempted by Ratiometric analysis. For ratiometric analysis, spectral peaks corresponding to the following major vibrational modes of fatty acids and esters were systematically identified; for example, Carbonyl stretching of carboxyl groups (\u0026ndash;COOH) vibration at 1742 cm⁻\u0026sup1;, methylene bending (δ-CH₂) at 1434 cm⁻\u0026sup1;, C-C skeletal vibrations at 1123 cm⁻\u0026sup1; and 1296 cm⁻\u0026sup1;, C-H stretching at 2720 cm⁻\u0026sup1;, and -CH\u003csub\u003e2\u003c/sub\u003e stretching at 3010 cm⁻\u0026sup1;. Unsaturated fatty acids exhibited additional characteristic bands at 1259 cm⁻\u0026sup1; (=\u0026thinsp;C\u0026ndash;H bending vibrations) and 1652 cm⁻\u0026sup1; (C\u0026thinsp;=\u0026thinsp;C stretching vibrations), with peak intensities directly proportional to the degree of unsaturation. All possible pairwise intensity ratios (Ratiometric markers) were computed using these key spectral bands. Optimal ratiometric markers were identified using statistical analysis and validation, as described below.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eStatistical selection, marker validation, and multivariate modeling\u003c/h2\u003e \u003cdiv id=\"Sec9\" class=\"Section3\"\u003e \u003ch2\u003eStatistical identification of ratio markers\u003c/h2\u003e \u003cp\u003eOptimization and validation of the ratiometric markers were carried out following the analytical workflow shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e4\u003c/span\u003e. All analyses were performed using open source Python package 3.12.9. After constructing the comprehensive set of ratiometric markers, a two-stage selection strategy were employed to identify the most discriminatory markers. First, a Random Forest (RF) classifier was applied to rank all ratiometric markers based on their marker importance scores. Random Forest creates an ensemble of decision trees trained on randomized data subsets and assigns feature importance based on their contribution to classification accuracy across the forest (Menze et al. \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; Shehata et al. \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). The RF model provided an independent, model-based assessment of each marker. It should be noted that feature importance rankings obtained from Random Forest models may be influenced by feature correlations and dataset characteristics and should therefore be interpreted in conjunction with complementary statistical analyses. The markers exhibiting the highest importance scores were chosen as the most effective markers and selected for further statistical validation using the Kruskal-Wallis test (Kaur et al. \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Pereira et al. \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), a non-parametric method for assessing group-wise differences. The ratios demonstrating statistically significant variations across oil classes (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05) in the Kruskal-Wallis analysis were selected for further screening and validation through one-way analysis of variance (ANOVA) (Davari et al. \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) to confirm discriminatory power. This validation step was further aided by coefficient of variance (CV) analysis to assess marker reproducibility and robustness. Ultimately, Principal Component Analysis (PCA) (Jolliffe and Cadima \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2016\u003c/span\u003e) was performed on the final set of the ratiometric markers thus obtained to assess their effectiveness in clustering different oil classes in multivariate space. The variance explained by each principal component were calculated, and score plots were generated to visualize clustering. To check whether the clustering observed in PCA is statistically significant, MANOVA (Zhu et al. \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) was performed on the ratios. Four complementary test statistics were computed: Wilks' lambda (Λ), Pillai's trace (V), Hotelling\u0026ndash;Lawley trace (T), and Roy's greatest root (Θ) ( Alkarkhi and Alqaraghuli \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) to confirm the robustness of multivariate separation. The marker selection and validation procedures were designed to be systematic and statistically grounded within the available dataset and within the scope of the present proof-of-concept study. Nevertheless, further evaluation using independent validation datasets will be valuable for confirming long-term marker stability and generalizability under broader application scenarios.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Results and discussion","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eInterpretation of Raman spectral peaks related to fatty acids and esters\u003c/h2\u003e \u003cp\u003eSpectral analysis was performed on the baseline-corrected, smoothened and normalized Raman spectra of five oil types and their corresponding chip samples (Fig.\u0026nbsp;5). Both oils and chips samples were found to contain a few strong peaks at 1123 cm⁻\u0026sup1; [(C\u0026ndash;C) stretching of the (CH2)\u003csub\u003en\u003c/sub\u003e group], 1259 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e (=\u0026thinsp;C\u0026ndash;H bending), 1296 cm⁻\u0026sup1; [(=\u0026thinsp;C\u0026ndash;H) deformation of cis(R\u0026ndash;HC\u0026thinsp;=\u0026thinsp;CH\u0026ndash;R)], 1434 cm⁻\u0026sup1; (methylene bending (δ-CH₂)), 1652 cm⁻\u0026sup1; [(C\u0026thinsp;=\u0026thinsp;C) of cis(R-HC\u0026thinsp;=\u0026thinsp;CH-R)], 1742 cm⁻\u0026sup1; [carbonyl stretching of ester carboxyl groups (-COOH)], 2720 cm⁻\u0026sup1; (C\u0026ndash;H stretching) and 3010 cm⁻\u0026sup1; (methyl group stretching). As mentioned in materials and methods section, the chips spectra were collected after blotting the oil from the chips using tissue paper. Therefore, presence of similar signature peaks in chips and oils suggests that the dominant molecular signatures of the oils are largely preserved under the proposed sampling conditions, despite effects from both the food matrix and the tissue paper (cellulose) matrix. It suggests tissue paper based sampling to be an effective alternative for tedious extraction based sampling which are most commonly used during food ingredient analysis. Overall, the preliminary spectral analysis of the data acquired from the oil laden tissue paper confirmed that Raman spectroscopy captured chemically meaningful lipid signatures in both oil and chips, paving way for a robust physicochemical foundation for subsequent ratiometric analysis and multivariate discrimination.\u003c/p\u003e \u003cp\u003eAfter assigning the peaks to specific vibrational groups, the peak intensities were systematically evaluated to determine their correlation with variations in saturated and unsaturated fatty acid composition across the studied oils. CV analysis was done on the intensity of the above mentioned peaks. Unfortunately, CV obtained, as low as 0.2\u0026ndash;0.3%, indicated their incapability to be used as markers. In order to find suitable markers, instead of relying on the individual peak intensities, intensity ratios of two chemically meaningful peaks (ratiometric markers) were explored.\u003c/p\u003e \u003cp\u003e \u003cb\u003eFigure\u0026nbsp;5\u003c/b\u003e Baseline corrected, smoothened and normalized Raman spectra collected from oils and chips\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eIdentifying ratiometric markers suitable for oil type classification\u003c/h2\u003e \u003cp\u003eTo identify the suitable ratiometric markers from the extensive pool of all potential markers, three crucial criteria were evaluated; chemical relevance, model-based marker-ranking approach and statistical significance. Only those ratios were retained which (i) were constructed from chemically interpretable Raman bands, (ii) showed high significance in Random forest markers importance and (iii) statistically significant as per Kruskal\u0026ndash;Wallis analysis. Ratios failing to meet these criteria were excluded.\u003c/p\u003e \u003cp\u003eRandom Forest (RF) classifier was employed on the chemically relevant markers as an independent, model-based marker-ranking approach. According to the RF model (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e), five ratiometric markers I₁₆₅₂/₁₇₄₂, I₁₆₅₂/₁₄₃₄, I₁₇₄₂/₁₂₅₉, I₁₆₅₂/₁₂₅₉, and I₁₂₉₆/₁₄₃₄ were identified as the most informative. Their importance values ranged from ~\u0026thinsp;0.300 to 0.116, while all other ratios showed much lower contributions. The statistical significance (p\u0026thinsp;\u0026lt;\u0026thinsp;0.005) of these highly ranked ratios were further assessed through Kruskal-Wallis group-wise testing, which revealed the same ratios to consistently show highly significant differences (p\u0026thinsp;\u0026lt;\u0026thinsp;\u0026lt;\u0026thinsp;10⁻⁶) across oil classes. This systematic selection resulted in a compact, chemically significant ratiometric marker set, which formed the basis for all subsequent multivariate classification.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eStatistical Significance by Kruskal\u0026ndash;Wallis analysis and Random Forest Importance of Ratiometric markers in Oils and Chips\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRatio markers\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eKruskal\u0026ndash;Wallis analysis\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003eRF importance\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOils\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eChips\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eOils\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eChips\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003ep value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003ep value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eI\u003csub\u003e1652/1434\u003c/sub\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;1 x 10\u003csup\u003e\u0026minus;\u0026thinsp;22\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;1 x 10\u003csup\u003e\u0026minus;\u0026thinsp;15\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.300\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.417\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eI\u003csub\u003e1296/1434\u003c/sub\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;1 x 10\u003csup\u003e\u0026minus;\u0026thinsp;6\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;1 x 10\u003csup\u003e\u0026minus;\u0026thinsp;9\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.116\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.139\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eI\u003csub\u003e1652/1742\u003c/sub\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;1 x 10\u003csup\u003e\u0026minus;\u0026thinsp;10\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;1 x 10\u003csup\u003e\u0026minus;\u0026thinsp;8\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.208\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.138\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eI\u003csub\u003e1742/1259\u003c/sub\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;1 x 10\u003csup\u003e16\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;1 x 10\u003csup\u003e\u0026minus;\u0026thinsp;13\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.212\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.131\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eI\u003csub\u003e1652/1259\u003c/sub\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;1 x 10\u003csup\u003e\u0026minus;\u0026thinsp;12\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;1 x 10\u003csup\u003e\u0026minus;\u0026thinsp;16\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.123\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.103\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eStatistical screening and reproducibility validation of ratiometric markers\u003c/h2\u003e \u003cp\u003eA cross-validated marker selection strategy was employed to systematically validate the ratiometric markers identified in previous section. Based on the ANOVA results (F-statistic) and CV values (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e), all five ratiometric markers: I₁₆₅₂/I₁₄₃₄, I₁₇₄₂/I₁₂₅₉, and I₁₆₅₂/I₁₇₄₂, I\u003csub\u003e1652/1259\u003c/sub\u003e and I\u003csub\u003e1296/1434\u003c/sub\u003e were identified as suitable for optimal oil type classification. For oil samples, these ratios exhibited strong statistical separation across oil types (F: 22629.79, 1267.22, and 11697.20 and CV: 43.77%, 43.86% and 38.23%, respectively). However, for chips samples, the F-statistics and CV values turned out to be lower (F: 2568.23, 331.46 and 282.45; CV: 43.64%, 38.82% and 33.09%). This reduction is most likely attributable to the increased heterogeneity and compositional complexity of the chips (food matrix). Overall, the results from the ratio markers analysis demonstrated that oil type classification can be potentially achieved using this identified and statistically significant set of five robust, chemically interpretable ratio markers and suitable for chemometric analysis for the final framework development.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eStatistical Significance of selective ratiometric markers assessed by ANOVA (F_statistic) and Coefficient of Variance (CV) values for oils and chips.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRatiometric Markers\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eOil Samples\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003eChips Samples\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eF_statistic\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCV Values (%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eF_statistic\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eCV Values (%)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eI\u003csub\u003e1652/1434\u003c/sub\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e22629.79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e43.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2568.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e43.64\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eI\u003csub\u003e1652/1742\u003c/sub\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e11697.20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e43.86\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e331.46\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e38.82\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eI\u003csub\u003e1742/1259\u003c/sub\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1267.22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e38.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e282.45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e33.09\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eI\u003csub\u003e1652/1259\u003c/sub\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e666.20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e34.52\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e55.76\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e10.58\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eI\u003csub\u003e1296/1434\u003c/sub\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e637.99\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e33.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e189.10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e20.48\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eMultivariate classification and statistical validation using optimized ratiometric markers\u003c/h2\u003e \u003cp\u003eThe classification efficiency of the selected ratio markers was further evaluated using multivariate analysis to assess whether oil types can form distinct clustering corresponding to different types of oils. Multivariate separation was explored using Principal Component Analysis (PCA), followed by formal statistical validation using MANOVA.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eOil data analysis\u003c/h2\u003e \u003cp\u003eAs seen in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e6\u003c/span\u003e, PCA of oil samples based on the selected ratiometric markers revealed discernible clustering, with the first two principal components explaining 76% and 15% of the total variance, respectively (91% cumulative variance). The grouping of oil types within the principal component space is evident in the corresponding score plot. Quantitative evaluation yielded a silhouette coefficient of 0.36 (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e), indicating weak-to-moderate cluster separation. In general silhouette coefficients range from \u0026minus;\u0026thinsp;1 to 0 to 1, with values close to zero reflecting overlapping clusters and values approaching unity indicating well-separated groups. In addition, the ratio of between-class to within-class scatter (B/W\u0026thinsp;=\u0026thinsp;3.34) demonstrated that inter-class variability substantially exceeded intra-class variability in the PCA space. A permutation test on the B/W statistic (p_perm\u0026thinsp;=\u0026thinsp;0.005) further confirmed that the observed clustering is unlikely to arise from random variation. Together, these complementary metrics indicate statistically meaningful separation between oil classes.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eChips data analysis\u003c/h2\u003e \u003cp\u003eFor chips, PCA revealed a reduced but still interpretable multivariate separation, with the first two principal components accounting for 57% and 20% of the variance, respectively (77% cumulative variance). Although cluster separation was less pronounced than in oils, quantitative measures still indicated non-random structure (silhouette\u0026thinsp;=\u0026thinsp;0.12; B/W\u0026thinsp;=\u0026thinsp;0.88; p_perm\u0026thinsp;=\u0026thinsp;0.005). This reduced separation is expected to result from food-matrix interference, yet oil-dependent trends remain detectable in multivariate space which is valuable from screening purpose.\u003c/p\u003e \u003cp\u003eWhile PCA and associated clustering metrics provide geometric and distributional measures of class separability, they do not constitute formal hypothesis tests. Therefore, multivariate analysis of variance (MANOVA) was employed to rigorously assess whether the observed group separation is statistically significant across all ratiometric variables simultaneously. Results obtained from the MANOVA analysis are presented in Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eEvaluation metrics by PCA and MANOVA\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colspan=\"7\" nameend=\"c7\" namest=\"c1\"\u003e \u003cp\u003ePCA metrics\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colspan=\"4\" nameend=\"c4\" namest=\"c1\"\u003e \u003cp\u003eOils\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"3\" nameend=\"c7\" namest=\"c5\"\u003e \u003cp\u003eChips\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003eFeatures\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eValues\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003eFeatures\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eValues\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSilhouette\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.361\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e\u003cb\u003eSilhouette\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.116\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eBetween_within_ratio\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3.33\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e\u003cb\u003eBetween_within_ratio\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.87\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003ep_perm\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.005\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e\u003cb\u003ep_perm\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.005\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"7\" nameend=\"c7\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMANOVA metrics\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c4\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eOils\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c7\" namest=\"c5\"\u003e \u003cp\u003e\u003cb\u003eChips\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eFeatures\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eValues\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003eF_value\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003ep\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003eValues\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003eF_value\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003ep\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eWilks' lambda\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.0008\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e984.83\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.0307\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e221.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003ePillai's trace\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2.2241\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e207.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.5068\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e89.94\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHotelling-Lawley trace\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e184.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e7609.35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e15.4828\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e573.41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRoy's greatest root\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e180.84\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e29929.62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e14.4794\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e2155.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eFor oils, Wilks\u0026rsquo; lambda indicated strong multivariate separation among oil classes (Λ\u0026thinsp;=\u0026thinsp;0.0008, F\u0026thinsp;=\u0026thinsp;984.83, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Wilks\u0026rsquo; lambda ranges from 0 to 1, with values close to unity indicating no group separation and values approaching zero reflecting near-complete discrimination. The observed value therefore indicates that only a negligible fraction of the total multivariate variance remains unexplained by oil type. Consistent results were obtained from Pillai\u0026rsquo;s trace (V\u0026thinsp;=\u0026thinsp;2.2241, F\u0026thinsp;=\u0026thinsp;207.28, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), which ranges from 0 to minimum (p, k\u0026thinsp;\u0026minus;\u0026thinsp;1), where (p\u0026thinsp;=\u0026thinsp;5) is the number of variables and (k\u0026thinsp;=\u0026thinsp;5) the number of groups, and quantifies the cumulative proportion of variance explained across multiple discriminant dimensions. In the present study, the observed Λ and V values correspond to more than half of the theoretical maximum, indicating strong multivariate separation. Similarly, the Hotelling\u0026ndash;Lawley trace (T\u0026thinsp;=\u0026thinsp;184.68, F\u0026thinsp;=\u0026thinsp;7609.35, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), which reflects the overall signal-to-noise ratio, and Roy\u0026rsquo;s greatest root (Θ\u0026thinsp;=\u0026thinsp;180.84, F\u0026thinsp;=\u0026thinsp;29929.62, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), which measures separation along the strongest discriminant axis, both indicate highly pronounced group differences. For chips, despite reduced discriminatory power due to food-matrix interference, significant multivariate separation was likewise observed (Λ\u0026thinsp;=\u0026thinsp;0.0307, V\u0026thinsp;=\u0026thinsp;1.5068, T\u0026thinsp;=\u0026thinsp;15.48, Θ\u0026thinsp;=\u0026thinsp;14.48; all p\u0026thinsp;\u0026lt;\u0026thinsp;0.0001). Although these values are smaller than those obtained for oils, they remain well within ranges associated with statistically meaningful discrimination. In particular, Wilks\u0026rsquo; lambda remains close to zero and Pillai\u0026rsquo;s trace indicates substantial explained variance, demonstrating that oil-dependent spectral signatures are retained despite interference from non-lipid components.\u003c/p\u003e \u003cp\u003eTaken together, the consistent significance across all four complementary statistics confirms robust and method-independent group separation. The close agreement between exploratory chemometric visualization (PCA) and formal multivariate inference (MANOVA) demonstrates that oil-type classification is driven by genuine multivariate oil-type-dependent compositional differences encoded in the ratiometric markers, establishing the statistical robustness and chemical relevance.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis study establishes a proof-of-concept Raman\u0026ndash;chemometric framework for rapid authentication of edible oils in processed foods using chemically interpretable ratiometric markers. Key markers were systematically identified through a two-stage selection strategy combining Random Forest importance ranking and Kruskal\u0026ndash;Wallis testing across all inter-peak combinations. Subsequent ANOVA and coefficient of variance (CV) analyses validated a final set of five robust markers exhibiting strong associations with saturated and unsaturated fatty acid content. PCA based chemometric analysis using these markers revealed pronounced clustering for oils (91% cumulative variance; silhouette\u0026thinsp;=\u0026thinsp;0.36; B/W\u0026thinsp;=\u0026thinsp;3.34) and moderate yet statistically significant separation for chips (77% cumulative variance; silhouette\u0026thinsp;=\u0026thinsp;0.12; B/W\u0026thinsp;=\u0026thinsp;0.88), with permutation testing confirming non-random structure (p\u0026thinsp;=\u0026thinsp;0.005). Multivariate analysis of variance further validated group separation for both oils (Wilks\u0026rsquo; Λ\u0026thinsp;=\u0026thinsp;0.0008, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and chips (Wilks\u0026rsquo; Λ\u0026thinsp;=\u0026thinsp;0.0307, p\u0026thinsp;\u0026lt;\u0026thinsp;0.0001), with consistent significance across complementary test statistics, demonstrating that the observed discrimination reflects genuine oil-type-dependent multivariate compositional differences rather than sampling variability.\u003c/p\u003e \u003cp\u003eIn parallel, this study introduces a novel tissue paper\u0026ndash;based minimal sampling strategy tailored for complex food matrices like fried chips, providing a practical and environmentally sustainable alternative to conventional solvent-intensive extraction protocols. This approach facilitates rapid, low-cost sample preparation while preserving analytical sensitivity in multi-component systems.\u003c/p\u003e \u003cp\u003eOverall, the proof-of-concept framework provides a foundation for automated and AI-assisted oil authentication workflow in processed foods. While the present study was conducted under controlled laboratory conditions using a limited number of oil types and a single food matrix, the robustness and chemical interpretability of the proposed framework provide a strong basis for extension to diverse industrial processing conditions, oil blends, varied adulteration scenarios, and commercial products. Future work will focus on expanding the methodology to a broader range of oil varieties and diverse food matrices. In addition, integration with advanced machine learning models will be explored to further enhance classification accuracy and enable scalable, real-time food authenticity monitoring and quality control.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003eThe authors would like to express their sincere gratitude to the Founder Chancellor and the Institute Management of Sri Sathya Sai Institute of Higher Learning, Anantapur, Andhra Pradesh, India, for their support in facilitating this research. The authors also acknowledge the Central Research Instrumentation Facility (CRIF) for providing the essential resources and facilities necessary for the instrumental analysis. Additionally, the authors extend their thanks to the Department of Food and Nutritional Sciences for their valuable resources and support throughout the course of this work.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contribution\u003c/strong\u003eJ.G. conceptualization, supervision, formal analysis, review \u0026amp; editing, approval of final version of manuscript. A.S. conceptualization, methodology, investigation, software, data curation, formal analysis, visualization, writing \u0026ndash; original draft, approval of final version of manuscript. C.S.N. software, data curation, formal analysis, visualization, review \u0026amp; editing, approval of final version of manuscript. S.M.V. supervision, review \u0026amp; editing, approval of final version of manuscript. \u0026nbsp;D.L.N.K. software, formal analysis, review \u0026amp; editing, approval of final version of manuscript. \u0026nbsp;B.P.R. data curation, formal analysis, approval of final version of manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability\u0026nbsp;\u003c/strong\u003eThe datasets and analysis scripts generated and used in the present study are available from the corresponding author upon reasonable request and will be shared on a case-by-case basis, subject to ongoing related research activities and institutional policies.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u0026nbsp;\u003c/strong\u003eThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u0026nbsp;\u003c/strong\u003eThe authors did not receive support from any organization for the submitted work.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAlkarkhi, A. F. M., Alqaraghuli, W. A. A. (2019). Comparing several population means. Easy Statistics for Food Science with R. Academic Press, pp 81-105. https://doi.org/10.1016/B978-0-12-814262-2.00006-6\u003c/li\u003e\n\u003cli\u003eAparicio R, Aparicio-Ruı́z R (2000) Authentication of vegetable oils by chromatographic techniques. Journal of Chromatography A 881:93\u0026ndash;104. https://doi.org/10.1016/S0021-9673(00)00355-1\u003c/li\u003e\n\u003cli\u003eAykas DP, Rodriguez-Saona LE (2016) Assessing potato chip oil quality using a portable infrared spectrometer combined with pattern recognition analysis. Anal Methods 8:731\u0026ndash;741. https://doi.org/10.1039/C5AY02387D\u003c/li\u003e\n\u003cli\u003eBaeten V, Hourant P, Morales MT, Aparicio R (1998) Oil and Fat Classification by FT-Raman Spectroscopy. J Agric Food Chem 46:2638\u0026ndash;2646. https://doi.org/10.1021/jf9707851\u003c/li\u003e\n\u003cli\u003eBansal G, Zhou W, Barlow PJ, et al (2010) Performance of palm olein in repeated deep frying and controlled heating processes. Food Chemistry 121:338\u0026ndash;347. https://doi.org/10.1016/j.foodchem.2009.12.034\u003c/li\u003e\n\u003cli\u003eBarton SJ, Ward TE, Hennelly BM (2018) Algorithm for optimal denoising of Raman spectra. Anal Methods 10:3759\u0026ndash;3769. https://doi.org/10.1039/C8AY01089G\u003c/li\u003e\n\u003cli\u003eBegum A, Jain BP (2024) Adulteration in edible oil (mustard oil) and ghee; detection and their effects on human health. Int J Biochem Mol Biol 15:141\u0026ndash;148. https://doi.org/10.62347/VVCI2092\u003c/li\u003e\n\u003cli\u003eComparing Several Population Means. https://www.sciencedirect.com/science/chapter/monograph/abs/pii/B9780128142622000066. Accessed 4 Feb 2026\u003c/li\u003e\n\u003cli\u003eCastro RC, Ribeiro DSM, Santos JLM, P\u0026aacute;scoa RNMJ (2022) The use of in-situ Raman spectroscopy to monitor at real time the quality of different types of edible oils under frying conditions. Food Control 136:108879. https://doi.org/10.1016/j.foodcont.2022.108879\u003c/li\u003e\n\u003cli\u003eCercaci L, Rodriguezestrada M, Lercker G (2003) Solid-phase extraction\u0026ndash;thin-layer chromatography\u0026ndash;gas chromatography method for the detection of hazelnut oil in olive oils by determination of esterified sterols. Journal of Chromatography A 985:211\u0026ndash;220. https://doi.org/10.1016/S0021-9673(02)01397-3\u003c/li\u003e\n\u003cli\u003eDavari M, Bahreini M, Sabzevari Z (2024) Developing a non-destructive method for the detection of egg quality and freshness using micro-Raman spectroscopy. Applied Food Research 4:100453. https://doi.org/10.1016/j.afres.2024.100453\u003c/li\u003e\n\u003cli\u003eDesk, T. L. (2024, June 6). How to identify fake cooking oil? (FSSAI recommended tips). The Times of India. https://timesofindia.indiatimes.com/life-style/health-fitness/health-news/how-to-identify-fake-cooking-oil-fssai-recommended-tips/articleshow/110694766.cms\u003c/li\u003e\n\u003cli\u003eDorni C, Sharma P, Saikia G, Longvah T (2018) Fatty acid profile of edible oils and fats consumed in India. Food Chemistry 238:9\u0026ndash;15. https://doi.org/10.1016/j.foodchem.2017.05.072\u003c/li\u003e\n\u003cli\u003eElumalai A, Natarajan V (2025) Advancements in analytical technologies for ensuring food quality and authentication: A comprehensive review. Journal of Food Composition and Analysis 139:107075. https://doi.org/10.1016/j.jfca.2024.107075\u003c/li\u003e\n\u003cli\u003eFatima S, Kumar V, Bhadauria G, Verma H (2023) Quality indicators based rapid test kits for detection of frying oil quality: A review. Food Chemistry Advances 2:100305. https://doi.org/10.1016/j.focha.2023.100305\u003c/li\u003e\n\u003cli\u003eForooghi E (2023) Applications of Raman spectroscopy in food safety and quality control. HHHM 4:2. https://doi.org/10.30502/jhhhm.2024.430203.1082\u003c/li\u003e\n\u003cli\u003eFoster R, Williamson CS, Lunn J (2009) BRIEFING PAPER: Culinary oils and their health effects. Nutrition Bulletin 34:4\u0026ndash;47. https://doi.org/10.1111/j.1467-3010.2008.01738.x\u003c/li\u003e\n\u003cli\u003eG\u0026oacute;mez-Caravaca AM, Maggio RM, Cerretani L (2016) Chemometric applications to assess quality and critical parameters of virgin and extra-virgin olive oil. A review. Analytica Chimica Acta 913:1\u0026ndash;21. https://doi.org/10.1016/j.aca.2016.01.025\u003c/li\u003e\n\u003cli\u003eGong W, Shi R, Chen M, et al (2019) Quantification and monitoring the heat-induced formation of trans fatty acids in edible oils by Raman Spectroscopy. Food Measure 13:2203\u0026ndash;2210. https://doi.org/10.1007/s11694-019-00140-5\u003c/li\u003e\n\u003cli\u003eGu H, Huang X, Sun Y, et al (2021) Intelligent evaluation of total polar compounds (TPC) content of frying oil based on fluorescence spectroscopy and low-field NMR. Food Chemistry 342:128242. https://doi.org/10.1016/j.foodchem.2020.128242\u003c/li\u003e\n\u003cli\u003eHaji A, Desalegn K, Hassen H (2023) Selected food items adulteration, their impacts on public health, and detection methods: A review. Food Sci Nutr 11:7534\u0026ndash;7545. https://doi.org/10.1002/fsn3.3732\u003c/li\u003e\n\u003cli\u003eHu R, He T, Zhang Z, et al (2019) Safety analysis of edible oil products via Raman spectroscopy. Talanta 191:324\u0026ndash;332. https://doi.org/10.1016/j.talanta.2018.08.074\u003c/li\u003e\n\u003cli\u003eHuang F, Li Y, Guo H, et al (2016) Identification of waste cooking oil and vegetable oil via Raman spectroscopy. J Raman Spectroscopy 47:860\u0026ndash;864. https://doi.org/10.1002/jrs.4895\u003c/li\u003e\n\u003cli\u003eHuang X, Li H, Ruan Y, et al (2022) An integrated approach utilizing raman spectroscopy and chemometrics for authentication and detection of adulteration of agarwood essential oils. Front Chem 10:1036082. https://doi.org/10.3389/fchem.2022.1036082\u003c/li\u003e\n\u003cli\u003eInteaz A (1989) Food Quality Assurance: Elements of Raw Materials Control. Canadian Institute of Food Science and Technology Journal 22:113\u0026ndash;115. https://doi.org/10.1016/S0315-5463(89)70337-0\u003c/li\u003e\n\u003cli\u003eJamieson LE, Li A, Faulds K, Graham D (2018a) Ratiometric analysis using Raman spectroscopy as a powerful predictor of structural properties of fatty acids. R Soc Open Sci 5:181483. https://doi.org/10.1098/rsos.181483\u003c/li\u003e\n\u003cli\u003eJamieson LE, Li A, Faulds K, Graham D (2018b) Ratiometric analysis using Raman spectroscopy as a powerful predictor of structural properties of fatty acids. R Soc Open Sci 5:181483. https://doi.org/10.1098/rsos.181483\u003c/li\u003e\n\u003cli\u003eJohnson DS, Saikia DN (2009) Fatty acids profile of Edible Oils and Fats in India\u003c/li\u003e\n\u003cli\u003eJolliffe IT, Cadima J (2016) Principal component analysis: a review and recent developments. Phil Trans R Soc A 374:20150202. https://doi.org/10.1098/rsta.2015.0202\u003c/li\u003e\n\u003cli\u003eKaimal A, Dhingra M, Singhal R (2021) Monitoring of oil quality from commercial fried foods‐A case study from India. Journal of Food Processing and Preservation 46:. https://doi.org/10.1111/jfpp.16138\u003c/li\u003e\n\u003cli\u003eKaur A, Scarborough P, Hieke S, et al (2016) The nutritional quality of foods carrying health-related claims in Germany, The Netherlands, Spain, Slovenia and the United Kingdom. European Journal of Clinical Nutrition 70:114. https://doi.org/10.1038/ejcn.2016.114\u003c/li\u003e\n\u003cli\u003eKhursheed M (2024) Chromatographic Techniques for the Detection and Identification of Olive Oil Adulteration. ReiDoCrea: Revista electr\u0026oacute;nica de investigaci\u0026oacute;n Docencia Creativa. https://doi.org/10.30827/Digibug.86578\u003c/li\u003e\n\u003cli\u003eKim N, Yu KS, Kim J, et al (2018) Chemical characteristics of potato chips fried in repeatedly used oils. Food Measure 12:1863\u0026ndash;1871. https://doi.org/10.1007/s11694-018-9800-x\u003c/li\u003e\n\u003cli\u003eKorepanov VI (2020) Asymmetric least-squares baseline algorithm with peak screening for automatic processing of the Raman spectra. Journal of Raman Spectroscopy 51:2061\u0026ndash;2065. https://doi.org/10.1002/jrs.5952\u003c/li\u003e\n\u003cli\u003eKwofie F, Lavine BK, Ottaway J, Booksh K (2020) Differentiation of Edible Oils by Type Using Raman Spectroscopy and Pattern Recognition Methods. Appl Spectrosc 74:645\u0026ndash;654. https://doi.org/10.1177/0003702819888220\u003c/li\u003e\n\u003cli\u003eLiu H, Chen Y, Shi C, et al (2020) FT-IR and Raman spectroscopy data fusion with chemometrics for simultaneous determination of chemical quality indices of edible oils during thermal oxidation. LWT 119:108906. https://doi.org/10.1016/j.lwt.2019.108906\u003c/li\u003e\n\u003cli\u003eMendes TO, Da Rocha RA, Porto BLS, et al (2015) Quantification of Extra-virgin Olive Oil Adulteration with Soybean Oil: a Comparative Study of NIR, MIR, and Raman Spectroscopy Associated with Chemometric Approaches. Food Anal Methods 8:2339\u0026ndash;2346. https://doi.org/10.1007/s12161-015-0121-y\u003c/li\u003e\n\u003cli\u003eMenze BH, Kelm BM, Masuch R, et al (2009) A comparison of random forest and its Gini importance with standard chemometric methods for the feature selection and classification of spectral data. BMC Bioinformatics 10:213. https://doi.org/10.1186/1471-2105-10-213\u003c/li\u003e\n\u003cli\u003eMoore JC, Spink J, Lipp M (2012) Development and Application of a Database of Food Ingredient Fraud and Economically Motivated Adulteration from 1980 to 2010. Journal of Food Science 77:R118\u0026ndash;R126. https://doi.org/10.1111/j.1750-3841.2012.02657.x\u003c/li\u003e\n\u003cli\u003eNavratilova K, Hurkova K, Hrbek V, Uttl L (2022) Metabolic fingerprinting strategy: Investigation of markers for the detection of extra virgin olive oil adulteration with soft-deodorized olive oils. Food Control 134: 108649. https://doi.org/10.1016/j.foodcont.2021.108649\u003c/li\u003e\n\u003cli\u003ePereira FJ, L\u0026oacute;pez R, Ferrer N, et al (2021) A comparative appraisal of Raman band ratioing and chemometric analysis for classification of ancient papyri. Journal of Cultural Heritage 52:55\u0026ndash;64. https://doi.org/10.1016/j.culher.2021.09.003\u003c/li\u003e\n\u003cli\u003eProgram HF (2024) Economically Motivated Adulteration (Food Fraud) Research Publications. FDA. https://www.fda.gov/food/compliance-enforcement-food/economically-motivated-adulteration food-fraud\u003c/li\u003e\n\u003cli\u003eRana R, Kumar N, Pandit A, et al (2014) Effect of various socio-economic factors on the consumption of processed potato products in Punjab. Indian Journal of Agricultural Marketing 28:24\u0026ndash;35. https://www.researchgate.net/profile/Rajesh-Rana 3/publication/264889204_Effect_of_various_socio economic_factors_on_the_consumption_of_processed_potato_products_in_Punjab/links/5771046 808ae6219474a3263/Effect-of-various-socio-economic-factors-on-the-consumption-of-processed potato-products-in-Punjab.pdf\u003c/li\u003e\n\u003cli\u003eRifna EJ, Pandiselvam R, Kothakota A, et al (2022) Advanced process analytical tools for identification of adulterants in edible oils \u0026ndash; A review. Food Chemistry 369:130898. https://doi.org/10.1016/j.foodchem.2021.130898\u003c/li\u003e\n\u003cli\u003eSharma R (2022). Food Safety and Standards Authority of India Press release, India. Retrieved from https.://www.fssai.gov.in/upload/press_release/2022/08/62fcb62aa2949Press_Release_Edible_oil_ 17_08_2022.pdf\u003c/li\u003e\n\u003cli\u003eShehata M, Dodd S, Mosca S, et al (2024) Application of Spatial Offset Raman Spectroscopy (SORS) and Machine Learning for Sugar Syrup Adulteration Detection in UK Honey. Foods 13:2425. https://doi.org/10.3390/foods13152425\u003c/li\u003e\n\u003cli\u003eSu R, Xu X, Wang X, et al (2011) Determination of organophosphorus pesticides in peanut oil by dispersive solid phase extraction gas chromatography\u0026ndash;mass spectrometry. Journal of Chromatography B 879:3423\u0026ndash;3428. https://doi.org/10.1016/j.jchromb.2011.09.016\u003c/li\u003e\n\u003cli\u003eTan C-H, Kong I, Irfan U, et al (2021) Edible Oils Adulteration: A Review on Regulatory Compliance and Its Detection Technologies. Journal of Oleo Science 70:1343\u0026ndash;1356. https://doi.org/10.5650/jos.ess21109\u003c/li\u003e\n\u003cli\u003e\u0026Uuml;n İ, OK S (2018) Analysis of olive oil for authentication and shelf life determination. J Food Sci Technol 55:2476\u0026ndash;2487. https://doi.org/10.1007/s13197-018-3165-3\u003c/li\u003e\n\u003cli\u003eVaskova H, Buckova M (2018) Multivariate study of Raman spectral data of edible oils. 14: 02022. https://doi.org/10.1051/matecconf/201712502022\u003c/li\u003e\n\u003cli\u003eWang J, Qian J, Xu M, et al (2025) Adulteration detection of multi-species vegetable oils in camellia oil using Raman spectroscopy: Comparison of chemometrics and deep learning methods. Food Chemistry 463:141314. https://doi.org/10.1016/j.foodchem.2024.141314\u003c/li\u003e\n\u003cli\u003eWu Y, Liu J, Xu R, et al (2024) Recent advances in ratiometric surface-enhanced Raman spectroscopy sensing strategies. Microchemical Journal 199:110127. https://doi.org/10.1016/j.microc.2024.110127\u003c/li\u003e\n\u003cli\u003eYadav S (2018) Edible oil adulterations: Current issues, detection techniques, and health hazards. Int J Chem Stud 6:1393\u0026ndash;1397\u003c/li\u003e\n\u003cli\u003eYang T, Zhao B, He L (2019) Raman instruments for food quality evaluation. In: Evaluation Technologies for Food Quality. Elsevier, pp 119\u0026ndash;143. https://www.chemijournal.com/archives/2018/vol6issue2/PartT/6-2-102-397.pdf\u003c/li\u003e\n\u003cli\u003eYao S, Aykas DP, Rodriguez-Saona L (2020) Rapid Authentication of Potato Chip Oil by Vibrational Spectroscopy Combined with Pattern Recognition Analysis. Foods 10:42. https://doi.org/10.3390/foods10010042\u003c/li\u003e\n\u003cli\u003eZhu P, Yang Q, Zhao H (2022) Identification of peanut oil origins based on Raman spectroscopy combined with multivariate data analysis methods. Journal of Integrative Agriculture 21:2777\u0026ndash;2785. https://doi.org/10.1016/j.jia.2022.07.026\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":true,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"food-analytical-methods","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"Learn more about [Food Analytical Methods](https://www.springer.com/journal/12161)","snPcode":"12161","submissionUrl":"https://submission.nature.com/new-submission/12161/3","title":"Food Analytical Methods","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Raman spectroscopy, Edible oil authentication, Ratiometric analysis, PCA, ANOVA, MANOVA","lastPublishedDoi":"10.21203/rs.3.rs-8872116/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8872116/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Authentication of cooking oils in processed foods is essential for food safety, quality control, and regulatory compliance, yet routine analysis remains constrained by solvent-intensive extraction procedures and limited applicability to complex food matrices. This study presents a proof-of-concept Raman\u0026ndash;chemometric framework for direct oil authentication in processed foods using minimal, solvent-free sample handling. Potato chips were selected as a representative fried food matrix, and five commonly used edible oils (sunflower, soybean, groundnut, palm, and vanaspati) were analyzed in both pure form and corresponding chip matrices to enable systematic cross-matrix evaluation. A one-step tissue paper blotting method was employed for rapid oil recovery, followed by Raman spectroscopic analysis. Fatty-acid-associated Raman bands were identified and systematically combined into chemically interpretable inter-peak intensity ratios reflecting variations in saturation and ester content. A statistically grounded two-stage marker selection workflow integrating Random Forest importance ranking with non-parametric Kruskal\u0026ndash;Wallis testing was applied, followed by using one-way analysis of variance and coefficient of variation analysis. Five robust ratiometric markers (I₁₆₅₂/₁₇₄₂, I₁₆₅₂/₁₄₃₄, I₁₇₄₂/₁₂₅₉, I₁₆₅₂/₁₂₅₉, and I₁₂₉₆/₁₄₃₄) showed strong association with saturated versus unsaturated fatty acid profiles. Multivariate analysis based on these markers revealed pronounced separation among oil types (91% explained variance; F-values up to 22,629; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and statistically significant discrimination within chip matrices (77% explained variance), despite attenuation effects from the food matrix. Multivariate analysis of variance confirmed robust separation (Wilks\u0026rsquo; Λ\u0026thinsp;=\u0026thinsp;0.0307; p\u0026thinsp;\u0026lt;\u0026thinsp;0.0001). Overall, this framework establishes an interpretable, extraction-free, and scalable foundation for high-throughput oil authenticity screening in processed foods.","manuscriptTitle":"Raman-Chemometric Framework for Rapid Authentication of Edible Oils in Processed Foods Using a Validated One Step Sampling Technique","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-24 12:33:15","doi":"10.21203/rs.3.rs-8872116/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-03-12T20:39:49+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-04T11:34:51+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-02-25T05:34:46+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"118025477258054164047355008108045283298","date":"2026-02-20T07:59:41+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"240427453240478777906038602819191344906","date":"2026-02-20T05:40:02+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-02-20T04:29:58+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-02-16T00:48:10+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-02-16T00:47:38+00:00","index":"","fulltext":""},{"type":"submitted","content":"Food Analytical Methods","date":"2026-02-13T12:40:05+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"food-analytical-methods","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"Learn more about [Food Analytical Methods](https://www.springer.com/journal/12161)","snPcode":"12161","submissionUrl":"https://submission.nature.com/new-submission/12161/3","title":"Food Analytical Methods","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"93e2e148-af16-4b86-adcb-378b47fa8e67","owner":[],"postedDate":"February 24th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-05-16T18:08:21+00:00","versionOfRecord":[],"versionCreatedAt":"2026-02-24 12:33:15","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8872116","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8872116","identity":"rs-8872116","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}