Application of near-infrared spectroscopy to predict the cooking times of aged common beans (Phaseolus vulgaris L.)
Introduction
Legumes are widely considered important foods as part of diverse, nutritious, healthy and environmentally sustainable diets (Willett et al., 2019; Castro-Guerrero et al., 2016). This is true for the majority of lower-income populations in developing regions that highly depend on plant foods due to the high cost of foods from animal sources (Chávez-Mendoza and Sánchez, 2017; Castro-Guerrero et al., 2016; Câmara et al., 2013) as well as for high-income populations (Willett et al., 2019). Among the legumes, common beans (Phaseolus vulgaris L.) are the most economically significant species and are widely accepted (Broughton et al., 2003). For all these reasons, common beans have great potential to contribute to food and nutrition security now and in the future.
The cooking process (cooking time) is one of the major factors affecting the consumption convenience and general acceptability of common beans (Katungi et al., 2011; Bernal-Lugo et al., 1997). Conventional cooking is usually a two-step procedure of texture degradation which starts with soaking followed by thermal processing, mostly boiling at household or industrial level. During cooking, the rate limiting step of bean softening has been reported to be the solubilization of pectin present in the middle lamella and the cell walls of the cotyledon (Chigwedere et al., 2018). The cooking process of beans has been previously measured subjectively and/or objectively as the percentage of beans that have attained a stipulated endpoint of texture degradation or chemical change (e.g. starch gelatinization), then called cooked. The time it takes to reach the desired endpoint is an important quality indicator for breeders, seed distributors, food industries and consumers (Wood, 2017).
The cooking time of beans depends on variety (Bernal-Lugo et al., 1997; Kinyanjui et al., 2015) and even within the same variety they can differ depending on postharvest storage conditions (Kinyanjui et al., 2017). Cooking times can range from 30 to 120 min when beans are freshly harvested (Garcia et al., 2012) up to 8 h due to inappropriate storage (Garcia et al., 1998). This has been attributed to the development of the Hard-To-Cook (HTC) textural defect. This defect causes the bean cotyledon to harden and although the bean properly hydrates, bean softening is hindered, resulting in a prolonged cooking time (Shiga and Lajolo, 2006; Shiga et al., 2004).
The mechanism of HTC development in beans has yet to be fully elucidated, nonetheless, several mechanisms have been suggested. The mechanisms depend on enzyme-catalyzed reactions and transport of small molecules largely controlled by the plasticizing effect of water. Therefore, temperature and relative humidity (typically above 25 °C and 65%) during storage (Stanley, 1992) and probably soaking are key factors of HTC development (Chigwedere et al., 2019a). The current overall hypothesis suggests that during adverse storage, membrane disruptions and (lipoxygenase catalyzed) lipid oxidation occur (Hussain et al., 1989). The pectin-cation-phytate theory (Kilmer et al., 1994; Hentges et al., 1991; Jones and Boulter, 1983), involves divalent cations released when phytase hydrolyzes phytates. The cations diffuse from protein bodies into the cell walls where they cross-link with pectin demethoxylated by pectin methylesterase. The lignification (Garcia et al., 1998; Hohlberg and Stanley, 1987) and multiple reactions theories (Aguilera and Rivera, 1992) involve the transport of phenolics to the cell walls resulting in lignification and/or phenolic-pectin cross-links. Besides the abovementioned theories, it has also been suggested that protein denaturation during storage and cooking hinders starch denaturation and thus bean softening (Garcia-Vela and Stanley, 1989).
To date, the approaches used to determine the cooking times of beans and eventually the HTC defect at the user level include finger pressing and texture analysis among others after cooking (Wood, 2017). Both these methods are time-consuming. Additionally, the finger pressing method requires trained personnel to obtain reliable results as it is subjective (Kinyanjui et al., 2015). This calls for further exploration of alternative techniques to address these challenges.
Near infra-red (NIR) spectroscopy is a high throughput analytical tool (Manley, 2014; Osborne, 2000b) that is widely applied in the quality check of raw materials and end-products in food research and industry (Nicolai et al., 2007; Williams, 2004; Osborne, 2000a). The NIR region of the electromagnetic spectrum ranges from 780 nm to 2800 nm. Samples are scanned to obtain spectra containing absorption information of functional groups (O–H, C–H, CO, N–H) and scattering information induced by microstructural heterogeneities. Subsequently, chemometrics is used to correlate the spectra to sample quality attributes obtained using conventional methods. This results in calibration models used to predict quality attributes of new samples (Saeys et al., 2019; Pasquini, 2018; Williams, 2004). NIR spectroscopy has been successfully used to predict the chemical composition, physical characteristics and sensory properties of common beans (Wang et al., 2014; Plans et al., 2014, 2013; 2012; Pojić et al., 2010).
NIR spectroscopy and hyperspectral imaging have also been successfully used to predict cooking times of Andean Diversity Panel accession covering a wide range of fresh common bean seed types (Mendoza et al., 2018b; Cichy et al., 2015). However, these studies excluded the development of storage-induced HTC. Also, the spectra were acquired for whole beans, which to the best of our knowledge likely put limitations on the vis/NIR approach when using the reflectance mode to evaluate cooking times associated with an interior structure, the cotyledon. Therefore, the aim of this study was to evaluate to what extent NIR spectroscopy can be used to determine the cooking time for aged beans. Hence, the experimental design involved using a limited number of varieties stored under different conditions to generate aged samples with a broad range of cooking times within each variety. As it has been shown that HTC development is cotyledon related, the aged samples were separated into two groups, the first with seed coats and the second as cotyledons only. The samples were milled before the acquisition of spectra. By doing this, we hypothesized that the molecules described in one or more of the mechanisms leading to HTC development could be identified during model development.
Section snippets
Materials and methods
The flow of research activities undertaken during the study is schematically presented in Fig. 1. Kinyanjui et al. (2017) conducted the storage and cooking experiments summarized in 2.1–2.2.
Logistic regression modelling of cooking kinetic data
Logistic regression modelling resulted in cooking times as summarized in Table 2. The corresponding cooking curves (experimental data points and predicted lines) are illustrated in Appendix A Figure A.1. The curves had an initial lag phase followed by an exponential phase and a final stationary phase. The duration of each phase was largely dependent on the variety, storage conditions and storage time. For all varieties the logistic regression predictor curves of samples stored for 0 months and
Conclusions
This study investigated the potential of NIR spectroscopy to predict cooking times of four varieties (Canadian wonder, pinto, red haricot and rose coco) of aged common beans. Firstly, evaluating cooking of beans using finger pressing followed by analyzing the kinetic data with logistic regression seems to be a suitable methodology to obtain cooking times. Secondly, these cooking times of whole aged beans can be predicted from the NIR spectra of milled raw beans. NIR spectra from milled
CRediT authorship contribution statement
Elizabeth Nakhungu Wafula: Conceptualization, Methodology, Investigation, Formal analysis, Visualization, Writing - original draft. Irene Njoki Wainaina: Conceptualization, Methodology, Investigation, Formal analysis, Visualization, Writing - review & editing. Carolien Buvé: Conceptualization, Methodology, Investigation, Formal analysis, Visualization, Writing - review & editing. Nghia-Do-Trong Nguyen: Conceptualization, Methodology, Investigation, Formal analysis, Visualization, Writing -
Declaration of competing interest
Please check the following as appropriate:
All authors have participated in (a) conception and design, or analysis and interpretation of the data; (b) drafting the article or revising it critically for important intellectual content; and (c) approval of the final version.
This manuscript has not been submitted to, nor is under review at, another journal or other publishing venue.
The authors have no affiliation with any organization with a direct or indirect financial interest in the subject
Acknowledgments
“The authors acknowledge the financial support from VLIR-UOS through funding of the Jomo Kenyatta University of Agriculture and Technology Institutional University Cooperation (IUC) programme ‘Legume Centre of Excellence for Food and Nutrition Security (LCEFoNS)’, grant number KE2017IUC037A101. Opinions of the author(s) do not automatically reflect those of either the Belgian government or VLIR-UOS, and can neither bind the Belgian Government nor VLIR-UOS.”
The funding source played no role in
References (56)
- et al.
Hard-to-cook defect in black beans : hardening rates, water imbibition and multiple mechanism hypothesis
Food Res. Int.
(1992) - et al.
Application of Fourier transform infrared spectroscopy to legume seed flour analysis
Food Chem.
(2008) - et al.
Mechanistic insight into softening of Canadian wonder common beans (Phaseolus vulgaris) during cooking
Food Res. Int.
(2018) - et al.
Insight into the evolution of fl avor compounds during cooking of common beans utilizing a headspace untargeted fi ngerprinting approach
Food Chem.
(2019) - et al.
Non-destructive measurement of fruit and vegetable quality by means of NIR spectroscopy: a review
Postharvest Biol. Technol.
(2007) Near infrared spectroscopy: a mature analytical technique with new perspectives – a review
Anal. Chim. Acta
(2018)- et al.
Estimating sensory properties of common beans (Phaseolus vulgaris L.) by near infrared spectroscopy
Food Res. Int.
(2014) - et al.
Characterization of common beans (Phaseolus vulgaris L.) by infrared spectroscopy: comparison of MIR, FT-NIR and dispersive NIR using portable and benchtop instruments
Food Res. Int.
(2013) - et al.
The development of near-infrared spectroscopy (NIRS) calibration for prediction of ash content in legumes on the basis of two different reference methods
Food Chem.
(2010) - et al.
Review of the most common pre-processing techniques for near-infrared spectra
Trac. Trends Anal. Chem.
(2009)
Comparative study of the composition and thermal properties of protein isolates prepared from nine Phaseolus vulgaris legume varieties
Food Res. Int.
Cell wall polysaccharides of common beans (Phaseolus vulgaris L.) - composition and structure
Carbohydr. Polym.
Changes in the cell wall polysaccharides during storage and hardening of beans
Food Chem.
A possible role for condensed tannins in bean hardening
Food Res. Int.
Near-infrared spectroscopy (NIRS) evaluation and regional analysis of Chinese faba bean (Vicia faba L.)
Crop J.
The lancet commissions food in the Anthropocene : the EAT – lancet commission on healthy diets from sustainable food systems
Lancet, February
Variable selection in regression — a tutorial
J. Chemometr.
Cotyledon thermal behavior and pectic solubility as related to cooking quality in common beans
Plant Foods Hum. Nutr.
Structural characteristics of stored black beans (Phaseolus vulgaris L.)
Scanning
Spp.) – model food legumes
Plant Soil
Pinto beans (Phaseolus vulgaris L.) as a functional food: implications on human health
Agriculture
Relationship between digestibility and secondary structure of raw and thermally treated legume proteins : a Fourier transform infrared (FT-IR) spectroscopic study
Amino Acids
Common bean: a legume model on the rise for unraveling responses and adaptations to iron, zinc, and phosphate deficiencies
Front. Plant Sci.
Bioactive compounds from mexican varieties of the common bean (Phaseolus vulgaris): implications for health
Molecules
Understanding the relations among the storage , soaking , and cooking behavior of Pulses : a scientific basis for innovations in sustainable foods for the future
Rev. Food Sci. Food Saf.
Genetic diversity and genome-wide association analysis of cooking time in dry bean (Phaseolus vulgaris L.)
Theor. Appl. Genet.
Multivariate Calibration: A Practical Guide for Developing Methods in the Quantitative Analytical Chemistry
Water-holding capacity in hard-to-cook beans (Phaseolus vulgaris): effect of pH and ionic strength
J. Food Sci.
Cited by (14)
Rapid and non-destructive detection of hard to cook chickpeas using NIR hyperspectral imaging and machine learning
2023, Food and Bioproducts ProcessingPaddy moisture on-line detection based on ensemble preprocessing and modeling for combine harvester
2022, Computers and Electronics in AgricultureCitation Excerpt :Other researchers have also reported similar absorption bands, and they found the absorption peaks of MC were located at 975 and 1450 nm, belonging to 950–1000 nm and 1430-1480 nm (Heman and Hsieh, 2016). A similar range of MC absorption bands was also observed in other grains of wheat (Shi and Yu, 2017), corn (Wang et al., 2020) and bean (Wafula et al., 2020). Meanwhile, the higher MC trend showed a higher absorbance, which can be found in the simplified enlarged version of 1150–1300 nm in Fig. 2.
Assessment of macronutrients and alpha-galactosides of texturized vegetable proteins by near infrared hyperspectral imaging
2022, Journal of Food Composition and AnalysisCitation Excerpt :According to other authors (Westad et al., 2013), the RPD values obtained ranged the models from being considered fair models (ashes) to very good models (carbohydrates). The obtained models showed good stability, as evidenced by the small differences between the RMSECV and the RMSEP (Wafula et al., 2020), with the latter being slightly lower than the former in several cases. Comparing our results with the outcomes obtained for other food products, the developed models are very promising for real application and better than those obtained, for example, in meat (Kamruzzaman et al., 2012).
Application of multivariate data analysis for food quality investigations: An example-based review
2022, Food Research International