FoodPro: A Web-Based Tool for Evaluating Covariance and Correlation NMR Spectra Associated with Food Processes.

Eisuke Chikayama,Jun Kikuchi,Ryo Yamashina,Keiko Komatsu,Yasuyo Sekiyama,Yuuri Tsuboi,Kenji Sakata

doi:10.3390/metabo6040036

Abstract

Foods from agriculture and fishery products are processed using various technologies. Molecular mixture analysis during food processing has the potential to help us understand the molecular mechanisms involved, thus enabling better cooking of the analyzed foods. To date, there has been no web-based tool focusing on accumulating Nuclear Magnetic Resonance (NMR) spectra from various types of food processing. Therefore, we have developed a novel web-based tool, FoodPro, that includes a food NMR spectrum database and computes covariance and correlation spectra to tasting and hardness. As a result, FoodPro has accumulated 236 aqueous (extracted in D2O) and 131 hydrophobic (extracted in CDCl3) experimental bench-top 60-MHz NMR spectra, 1753 tastings scored by volunteers, and 139 hardness measurements recorded by a penetrometer, all placed into a core database. The database content was roughly classified into fish and vegetable groups from the viewpoint of different spectrum patterns. FoodPro can query a user food NMR spectrum, search similar NMR spectra with a specified similarity threshold, and then compute estimated tasting and hardness, covariance, and correlation spectra to tasting and hardness. Querying fish spectra exemplified specific covariance spectra to tasting and hardness, giving positive covariance for tasting at 1.31 ppm for lactate and 3.47 ppm for glucose and a positive covariance for hardness at 3.26 ppm for trimethylamine N-oxide.

Highlights

Foods from agriculture and fishery products are processed using various technologies.A given foodstuff is a complex matrix including several compounds, from small molecules to macromolecules and with a wide range of physicochemical properties
As nuclear magnetic resonance (NMR) has great potential to characterize molecular mixtures [7,8], we have focused on realizing its potential for food analysis [9,10,11,12]
Nuclear Magnetic Resonance (NMR) approaches for analyzing harvested raw agriculture or fishery materials have been widely reported by us [16,17,18,19,20] and several other groups [21,22,23], systematic analysis throughout food processing toward a variety of materials is not reported, to the best of our knowledge

Summary

Introduction

Foods from agriculture and fishery products are processed using various technologies. According to cross-site analytical validity studies, the interconvertibility of NMR data among different institutions is a distinct advantage of the NMR-based approach [13,14,15]. Such wide data dissemination is essential for industrial applications such as evaluation of agricultural products and food quality control. NMR approaches for analyzing harvested raw agriculture or fishery materials have been widely reported by us [16,17,18,19,20] and several other groups [21,22,23], systematic analysis throughout food processing toward a variety of materials is not reported, to the best of our knowledge. Our developed tool aimed to predict similar spectral/physical/subjective properties when querying with a user NMR spectrum

Results and Discussion

Typical experimentalbench-top bench-top 60-MHz

Summary

Accumulation of Experimental Data

Database Design

Computation of Covariance and Correlation Spectra for Tasting and Hardness