Predicting odor profile of food from its chemical composition: Towards an approach based on artificial intelligence and flavorists expertise

N. Mejean Perrot; Alice Roche; Alberto Tonda; Evelyne Lutton; Thierry Thomas-Danguin; N. Mejean Perrot; Alice Roche; Alberto Tonda; Evelyne Lutton; Thierry Thomas-Danguin

doi:10.3934/mbe.2023908

Mathematical Biosciences and Engineering

2023, Volume 20, Issue 12: 20528-20552. doi: 10.3934/mbe.2023908

Previous Article Next Article

Research article Special Issues

Predicting odor profile of food from its chemical composition: Towards an approach based on artificial intelligence and flavorists expertise

1.
UMR 518 MIA-PS, INRAE, AgroParisTech, Université Paris-Saclay, 22 place de l'Agronomie, 91120, Palaiseau, France
2.
Centre des Sciences du Goût et de l'Alimentation, INRAE, CNRS, Institut Agro Dijon, Université de Bourgogne Franche-Comté, F-21000 Dijon, France
3.
Institut des Systèmes Complexes de Paris Île-de-France (ISC-PIF), UAR 3611 CNRS, 75013 Paris, France
^‡The two authors share first authorship

Received: 31 May 2023 Revised: 23 October 2023 Accepted: 30 October 2023 Published: 14 November 2023

Odor is central to food quality. Still, a major challenge is to understand how the odorants present in a given food contribute to its specific odor profile, and how to predict this olfactory outcome from the chemical composition. In this proof-of-concept study, we seek to develop an integrative model that combines expert knowledge, fuzzy logic, and machine learning to predict the quantitative odor description of complex mixtures of odorants. The model output is the intensity of relevant odor sensory attributes calculated on the basis of the content in odor-active comounds. The core of the model is the mathematically formalized knowledge of four senior flavorists, which provided a set of optimized rules describing the sensory-relevant combinations of odor qualities the experts have in mind to elaborate the target odor sensory attributes. The model first queries analytical and sensory databases in order to standardize, homogenize, and quantitatively code the odor descriptors of the odorants. Then the standardized odor descriptors are translated into a limited number of odor qualities used by the experts thanks to an ontology. A third step consists of aggregating all the information in terms of odor qualities across all the odorants found in a given product. The final step is a set of knowledge-based fuzzy membership functions representing the flavorist expertise and ensuring the prediction of the intensity of the target odor sensory descriptors on the basis of the products' aggregated odor qualities; several methods of optimization of the fuzzy membership functions have been tested. Finally, the model was applied to predict the odor profile of 16 red wines from two grape varieties for which the content in odorants was available. The results showed that the model can predict the perceptual outcome of food odor with a certain level of accuracy, and may also provide insights into combinations of odorants not mentioned by the experts.
- Aroma,
- machine learning,
- expert knowledge,
- Fuzzy logic,
- sensory prediction,
- inclusive model
Citation: N. Mejean Perrot, Alice Roche, Alberto Tonda, Evelyne Lutton, Thierry Thomas-Danguin. Predicting odor profile of food from its chemical composition: Towards an approach based on artificial intelligence and flavorists expertise[J]. Mathematical Biosciences and Engineering, 2023, 20(12): 20528-20552. doi: 10.3934/mbe.2023908

Related Papers:

Abstract

Odor is central to food quality. Still, a major challenge is to understand how the odorants present in a given food contribute to its specific odor profile, and how to predict this olfactory outcome from the chemical composition. In this proof-of-concept study, we seek to develop an integrative model that combines expert knowledge, fuzzy logic, and machine learning to predict the quantitative odor description of complex mixtures of odorants. The model output is the intensity of relevant odor sensory attributes calculated on the basis of the content in odor-active comounds. The core of the model is the mathematically formalized knowledge of four senior flavorists, which provided a set of optimized rules describing the sensory-relevant combinations of odor qualities the experts have in mind to elaborate the target odor sensory attributes. The model first queries analytical and sensory databases in order to standardize, homogenize, and quantitatively code the odor descriptors of the odorants. Then the standardized odor descriptors are translated into a limited number of odor qualities used by the experts thanks to an ontology. A third step consists of aggregating all the information in terms of odor qualities across all the odorants found in a given product. The final step is a set of knowledge-based fuzzy membership functions representing the flavorist expertise and ensuring the prediction of the intensity of the target odor sensory descriptors on the basis of the products' aggregated odor qualities; several methods of optimization of the fuzzy membership functions have been tested. Finally, the model was applied to predict the odor profile of 16 red wines from two grape varieties for which the content in odorants was available. The results showed that the model can predict the perceptual outcome of food odor with a certain level of accuracy, and may also provide insights into combinations of odorants not mentioned by the experts.

References

[1]	D. E. Hornung, M. P. Enns, The contributions of smell and taste to overall intensity: A model, Percept. Psychophys., 39 (1986), 385–391. https://doi.org/10.3758/BF03207066 doi: 10.3758/BF03207066
[2]	B. Lorrain, J. Ballester, T. Thomas-Danguin, J. Blanquet, J. M. Meunier, Y. Le Fur, Selection of potential impact odorants and sensory validation of their importance in typical chardonnay wines, J. Agri. Food Chem., 54 (2006), 3973–3981.
[3]	C. Flavián, S. Ibáñez-Sánchez, C. Orús, The influence of scent on virtual reality experiences: The role of aroma-content congruence, J. Business Res., 123 (2021), 289–301. https://doi.org/10.1016/j.jbusres.2020.09.036 doi: 10.1016/j.jbusres.2020.09.036
[4]	R. Hudson, Editorial note, Chem. Senses, 25 (2000), 693–693. https://doi.org/10.1046/j.1442-9993.2000.00002.x doi: 10.1046/j.1442-9993.2000.00002.x
[5]	H. Stone, R. N. Bleibaum, H. A. Thomas, Sensory evaluation practices, Academic press, 2012.
[6]	G. Ferrari, O. Lablanquie, R. Cantagrel, J. Ledauphin, T. Payot, N. Fournier, et al., Determination of key odorant compounds in freshly distilled cognac using GC-o, GC-MS, and sensory evaluation, J. Agri. Food Chem., 52 (2004), 5670–5676.
[7]	S.-J. Lee, A. C. Noble, Characterization of odor-active compounds in californian chardonnay wines using GC-olfactometry and GC-mass spectrometry, J. Agri. Food Chem., 51 (2003), 8036–8044.
[8]	A. Pérez-Silva, E. Odoux, P. Brat, F. Ribeyre, G. Rodriguez-Jimenes, V. Robles-Olvera, et al., GC–MS and GC–olfactometry analysis of aroma compounds in a representative organic aroma extract from cured vanilla (vanilla planifolia g. jackson) beans, Food Chem, 99 (2006), 728–735. https://doi.org/10.1016/j.foodchem.2005.08.050 doi: 10.1016/j.foodchem.2005.08.050
[9]	M. Brattoli, E. Cisternino, P. Dambruoso, G. de Gennaro, P. Giungato, A. Mazzone, et al., Gas chromatography analysis with olfactometric detection (GC-o) as a useful methodology for chemical characterization of odorous compounds, Sensors, 13 (2013), 16759–16800. https://doi.org/10.3390/s131216759 doi: 10.3390/s131216759
[10]	T. Thomas-Danguin, C. Sinding, S. Romagny, F. E. Mountassir, B. Atanasova, E. Le Berre, et al., The perception of odor objects in everyday life: A review on the processing of odor mixtures, Front. Psychol., 5 (2014).
[11]	W. Grosch, Evaluation of the key odorants of foods by dilution experiments, aroma models and omission, Chem. Senses, 26 (2001), 533–545. https://doi.org/10.1093/chemse/26.5.533 doi: 10.1093/chemse/26.5.533
[12]	H. Guth, Quantitation and sensory studies of character impact odorants of different white wine varieties, J. Agri. Food Chem., 45 (1997), 3027–3032.
[13]	V. Ferreira, N. Ortín, A. Escudero, R. López, J. Cacho, Chemical characterization of the aroma of grenache rosé wines: aroma extract dilution analysis, quantitative determination, and sensory reconstitution studies, J. Agri. Food Chem., 50 (2002), 4048–4054.
[14]	A. Escudero, B. Gogorza, M. A. Melús, N. Ortín, J. Cacho, V. Ferreira, Characterization of the aroma of a wine from maccabeo. key role played by compounds with low odor activity values, J. Agri. Food Chem., 52 (2004), 3516–3524.
[15]	V. Ferreira, M. Sáenz-Navajas, E. Campo, P. Herrero, A. de la Fuente, P. Fernández-Zurbano, Sensory interactions between six common aroma vectors explain four main red wine aroma nuances, Food Chem., 199 (2016), 447–456. https://doi.org/10.1016/j.foodchem.2015.12.048 doi: 10.1016/j.foodchem.2015.12.048
[16]	Y. Zheng, B. G. Sun, M. M. Zhao, F. P. Zheng, M. Q. Huang, J. Y. Sun, et al., Characterization of the key odorants in chinese zhima aroma-type baijiu by gas chromatography–olfactometry, quantitative measurements, aroma recombination, and omission studies, J. Agri. Food Chem., 64 (2016), 5367–5374.
[17]	W. S. Cain, Odor intensity: Differences in the exponent of the psychophysical function, Percept. Psychophys., 6 (1969), 349–354. https://doi.org/10.3758/BF03212789 doi: 10.3758/BF03212789
[18]	P. A. Edwards, P. C. Jurs, Correlation of odor intensities with structural properties of odorants, Chem. Senses, 14 (1989), 281–291. https://doi.org/10.1093/chemse/14.2.281 doi: 10.1093/chemse/14.2.281
[19]	M. H. Abraham, R. Sanchez-Moreno, J. E. Cometto-Muniz, W. S. Cain, An algorithm for 353 odor detection thresholds in humans, Chemical Senses, 37 (2012), 207–218. https://doi.org/10.1093/chemse/bjr094 doi: 10.1093/chemse/bjr094
[20]	A. Keller, R. C. Gerkin, Y. F. Guan, A. Dhurandhar, G. Turu, B. Szalai, et al., Predicting human olfactory perception from chemical features of odor molecules, Science, 355 (2017), 820–826. https://doi.org/10.1126/science.aal2014 doi: 10.1126/science.aal2014
[21]	R. M. Khan, C.-H. Luk, A. Flinker, A. Aggarwal, H. Lapid, R. Haddad, et al., Predicting odor pleasantness from odorant structure: Pleasantness as a reflection of the physical world, J. Neurosci., 27 (2007), 10015–10023. https://doi.org/10.1523/JNEUROSCI.1158-07.2007 doi: 10.1523/JNEUROSCI.1158-07.2007
[22]	A. Arshamian, R. C. Gerkin, N. Kruspe, E. Wnuk, S. Floyd, C. O'Meara, et al., The perception of odor pleasantness is shared across cultures, Current Biol., 32 (2022), 2061–2066. https://doi.org/10.1016/j.cub.2022.02.062 doi: 10.1016/j.cub.2022.02.062
[23]	L. Shang, C. J. Liu, Y. Tomiura, K. S. Hayashi, Machine-learning-based olfactometer: Prediction of odor perception from physicochemical features of odorant molecules, Anal. Chem., 89 (2017), 11999–12005. https://doi.org/10.1021/acs.analchem.7b02389 doi: 10.1021/acs.analchem.7b02389
[24]	K. Snitz, A. Yablonka, T. Weiss, I. Frumin, R. M. Khan, N. Sobel, Predicting odor perceptual similarity from odor structure, PLoS Comput. Biol., 9 (2013), e1003184. https://doi.org/10.1371/journal.pcbi.1003184 doi: 10.1371/journal.pcbi.1003184
[25]	A. C. Noble, S. E. Ebeler, Use of multivariate statistics in understanding wine flavor, Food Rev. Int., 18 (2002), 1–21. https://doi.org/10.1081/FRI-120003414 doi: 10.1081/FRI-120003414
[26]	M. Aznar, R. López, J. Cacho, V. Ferreira, Prediction of aged red wine aroma properties from aroma chemical composition. partial least squares regression models, J. Agri. Food Chem., 51 (2003), 2700–2707.
[27]	S. G. Penella, R. Boulanger, I. Maraval, G. Kopp, M. Corno, B.Fontez, et al., Link between flavor perception and volatile compound composition of dark chocolates derived from trinitario cocoa beans from dominican republic, Molecules, 28 (2023), 3805. https://doi.org/10.3390/molecules28093805 doi: 10.3390/molecules28093805
[28]	B Veinand, Sensory testing with flavourists: challenges and solutions, In Rapid Sensory Profiling Techniques, pages 383–399. Elsevier, 2015.
[29]	J.-N. Jaubert, C. Tapiero, J.-C. Dore, The field of odors: toward a universal language for odor relationships, Perfumer & Flavor., 20 (1995), 1–16.
[30]	S. Mirri, C. Prandi, M. Roccetti, P. Salomoni, Handmade Narrations: Handling digital narrations on food and gastronomic culture, J. Comput. Cultural Herit., 10 (2017), 1–17.
[31]	M. Sicard, C. Baudrit, M. N. Leclerc-Perlat, P. H. Wuillemin, N. Perrot, Expert knowledge integration to model complex food processes. application on the camembert cheese ripening process, Expert Syst. Appl., 38 (2011), 11804–11812.
[32]	N. Perrot, I. C. Trelea, C. Baudrit, G. Trystram, P. Bourgine, Modelling and analysis of complex food systems: State of the art and new trends, Trends Food Sci. Technol., 22 (2011), 304–314.
[33]	L. A. Zadeh, Fuzzy sets, Inform. Control, 8 (1965), 338–353. https://doi.org/10.1016/S0019-9958(65)90241-X
[34]	N. Perrot, I. Ioannou, I. Allais, C. Curt, J. Hossenlopp, G. Trystram, Fuzzy concepts applied to food product quality control: A review, Fuzzy Sets Syst., 157 (2006), 1145–1154. https://doi.org/10.1016/j.fss.2005.12.013 doi: 10.1016/j.fss.2005.12.013
[35]	J. Tan, X. Gao, D. E. Gerrard, application of fuzzy sets and neural networks in sensory analysis, J. Sensory Studies, 14 (1999), 119–138. https://doi.org/10.1111/j.1745-459X.1999.tb00108.x doi: 10.1111/j.1745-459X.1999.tb00108.x
[36]	V. J. Davidson, J. Ryks, T. Chu, Fuzzy models to predict consumer ratings for biscuits based on digital image features, IEEE Transact. Fuzzy Syst., 9 (2001), 62–67.
[37]	I. Ioannou, N. Perrot, J. Hossenlopp, G. Mauris, G. Trystram, The fuzzy set theory: a helpful tool for the estimation of sensory properties of crusting sausage appearance by a single expert, Food Quality Prefer., 13 (2002), 589–595. https://doi.org/10.1016/S0950-3293(02)00045-9 doi: 10.1016/S0950-3293(02)00045-9
[38]	S. Mukhopadhyay, G. C. Majumdar, T. K. Goswami, H. N. Mishra, Fuzzy logic (similarity analysis) approach for sensory evaluation of chhana podo, LWT - Food Sci. Technol., 53 (2013), 204–210.
[39]	C. Debjani, S. Das, H. Das, Aggregation of sensory data using fuzzy logic for sensory quality evaluation of food, J. Food Sci. Technol., 50 (2011), 1088–1096.
[40]	K. J. Shinde, I. L. Pardeshi, Fuzzy logic model for sensory evaluation of commercially available jam samples, J. Ready Eat Food, 1 (2014), 78–84.
[41]	O. Folorunso, Y. Ajayi, T. Shittu, Fuzzy-rule-based approach for modeling sensory acceptabitity of food products, Data Sci. J., 8 (2009), 70–77. https://doi.org/10.2481/dsj.007-006 doi: 10.2481/dsj.007-006
[42]	A. Villière, R. Symoneaux, A. Roche, A. Eslami, N. Perrot, Y. Le Fur, et al., Comprehensive sensory and chemical data on the flavor of 16 red wines from two varieties: sensory descriptive analysis, hs-spme-gc-ms volatile compounds quantitative analysis, and odor-active compounds identification by hs-spme-gc-ms-o, Data Brief, 24 (2019), 103725. https://doi.org/10.1016/j.dib.2019.103725 doi: 10.1016/j.dib.2019.103725
[43]	A. Loison, R. Symoneaux, P. Deneulin, T. Thomas-Danguin, C. Fant, L. Guérin, et al., Exemplarity measurement and estimation of the level of interjudge agreement for two categories of french red wines, Food Quality Prefer., 40 (2015), 240–251. https://doi.org/10.1016/j.foodqual.2014.10.001 doi: 10.1016/j.foodqual.2014.10.001
[44]	P. Pollien, A. Ott, F. Montigon, M. Baumgartner, R. Muñoz-Box, A. Chaintreau, Hyphenated headspace-gas chromatography-sniffing technique: Screening of impact odorants and quantitative aromagram comparisons, J. Agri. Food Chem., 45 (1997), 2630–2637.
[45]	S. Arctander, Perfume and flavor chemicals, Montlcair, N.J. USA, 1969.
[46]	W. Luebke, the good scents company, https://WWW.thegoodscentscompany.com/index.html, 1980.
[47]	L. A. Zadeh, Fuzzy logic and approximate reasoning, Synthese, 30 (1975), 407–428. https://doi.org/10.1007/BF00485052 doi: 10.1007/BF00485052
[48]	D. Dubois, H. Prade, Fuzzy sets and systems: Theory and applications, Academic press, NY, USA, 1969.
[49]	S. Arlot, A. Celisse, A survey of cross-validation procedures for model selection, Statistics Surveys, 4 (2010), 40–79.
[50]	W. Guang, M. Baraldo, M. Furlanut, Calculating percentage prediction error: A user's note, Pharmacolog. Res., 32 (1995), 241–248. https://doi.org/10.1016/S1043-6618(05)80029-5 doi: 10.1016/S1043-6618(05)80029-5
[51]	A. Roche, N. M. Perrot, T. Thomas-Danguin, Oops, the ontology for odor perceptual space: From molecular composition to sensory attributes of odor objects, Molecules, 2022. https://doi.org/10.3390/molecules272227888 doi: 10.3390/molecules272227888
[52]	N. Hansen, Towards a New Evolutionary Computation Studies in Fuzziness and Soft Computing, In chapter The CMA evolution strategy: A comparing review, (eds) J.A. Lozano, P. Larrañaga, I. Inza, I. Bengoetxea, 192 (2006), Springer, 75–102. https://doi.org/10.1080/02693445.1948.12035642

Reader Comments

Your name:*

Email:*
© 2023 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)