Biochemical Profile of Heritage and Modern Apple Cultivars and Application of Machine Learning Methods To Predict Usage, Age, and Harvest Season

FlavorMiner: a machine learning platform for extracting molecular flavor profiles from structural data

Flavor is the main factor driving consumers acceptance of food products. However, tracking the biochemistry of flavor is a formidable challenge due to the complexity of food composition. Current methodologies for linking individual molecules to flavor in foods and beverages are expensive and time-consuming. Predictive models based on machine learning (ML) are emerging as an alternative to speed up this process. Nonetheless, the optimal approach to predict flavor features of molecules remains elusive. In this work we present FlavorMiner, an ML-based multilabel flavor predictor. FlavorMiner seamlessly integrates different combinations of algorithms and mathematical representations, augmented with class balance strategies to address the inherent class of the input dataset. Notably, Random Forest and K-Nearest Neighbors combined with Extended Connectivity Fingerprint and RDKit molecular descriptors consistently outperform other combinations in most cases. Resampling strategies surpass weight balance methods in mitigating bias associated with class imbalance. FlavorMiner exhibits remarkable accuracy, with an average ROC AUC score of 0.88. This algorithm was used to analyze cocoa metabolomics data, unveiling its profound potential to help extract valuable insights from intricate food metabolomics data. FlavorMiner can be used for flavor mining in any food product, drawing from a diverse training dataset that spans over 934 distinct food products.Scientific Contribution FlavorMiner is an advanced machine learning (ML)-based tool designed to predict molecular flavor features with high accuracy and efficiency, addressing the complexity of food metabolomics. By leveraging robust algorithmic combinations paired with mathematical representations FlavorMiner achieves high predictive performance. Applied to cocoa metabolomics, FlavorMiner demonstrated its capacity to extract meaningful insights, showcasing its versatility for flavor analysis across diverse food products. This study underscores the transformative potential of ML in accelerating flavor biochemistry research, offering a scalable solution for the food and beverage industry.

Green Extraction of Bioactive Compounds from Apple Pomace from the Cider Industry

The cider-making industry in Asturias generates between 9000 and 12,000 tons of apple pomace per year. This by-product, the remains of the apple pressing, and made up of peel, flesh, seeds and stems, is a valuable material, containing substantial amounts of antioxidant compounds associated with healthy properties. Polyphenols such as dihydrochalcones and quercetin glycosides, and triterpenic acids, among which ursolic acid is a major compound, are the main antioxidant families described in apple pomace. The simultaneous recovery of those families has been accomplished by low frequency ultrasound-assisted extraction. Working extraction conditions were optimised by response surface methodology (RSM): time, 5.1 min; extractant composition, 68% ethanol in water; solid/liquid ratio, 1/75 and ultrasonic wave amplitude, 90%. This procedure was further applied to analyse those components in the whole apple pomace (WAP), apple peel (AP) and apple flesh (AF). On average, dry WAP contained almost 1300 µg/g of flavonols, 1200 µg/g of dihydrochalcones and 4200 µg/g of ursolic acid. These figures increased in the apple peel to, respectively 2500, 1400 and 8500 µg/g dry matter. Two linear multivariate regression models allowed the antioxidant activity of apple by-products to be predicted on the basis of their bioactive composition. The results derived from this study confirm the potential of industrial cider apple pomace as a source of high-value bioactive compounds, and the feasibility of the ultrasound-assisted extraction technique to recover those components in a simple and efficient way.

Machine Learning-Assisted Portable Microplasma Optical Emission Spectrometer for Food Safety Monitoring

To meet the needs of food safety for simple, rapid, and low-cost analytical methods, a portable device based on a point discharge microplasma optical emission spectrometer (μPD-OES) was combined with machine learning to enable on-site food freshness evaluation and detection of adulteration. The device was integrated with two modular injection units (i.e., headspace solid-phase microextraction and headspace purge) for the examination of various samples. Aromas from meat and coffee were first introduced to the portable device. The aroma molecules were excited to specific atomic and molecular fragments at excited states by room temperature and atmospheric pressure microplasma due to their different atoms and molecular structures. Subsequently, different aromatic molecules obtained their own specific molecular and atomic emission spectra. With the help of machine learning, the portable device was successfully applied to the assessment of meat freshness with accuracies of 96.0, 98.7, and 94.7% for beef, pork, and chicken meat, respectively, through optical emission patterns of the aroma at different storage times. Furthermore, the developed procedures can identify beef samples containing different amounts of duck meat with an accuracy of 99.5% and classify two coffee species without errors, demonstrating the great potential of their application in the discrimination of food adulteration. The combination of machine learning and μPD-OES provides a simple, portable, and cost-effective strategy for food aroma analysis, potentially addressing field monitoring of food safety.

Morphological Traits Evaluated with Random Forest Method Explains Natural Classification of Grapevine (Vitis vinifera L.) Cultivars

There are hundreds of morphologic and morphometric traits available to classify and identify grapevine (Vitis vinifera L.) genotypes, while statistical evaluation of those has certain limitations, especially when we have no information about the traits that are discriminative to a certain sample set. High numbers of investigated characters could cause redundancy, while reducing those numbers may result in data loss. Grapevine is one of the most important horticultural crops, with many cultivars in production. The characterization of the genotypes is of undeniably high importance. In this study, we analyzed a dataset of scientific and historical importance with 125 morphological traits of 97 grapevine cultivars described by Németh in 1966. However, the traits are not independent in a set of a large number of categorical traits with too few cultivars. Therefore, the number of traits was first reduced using a simple and effective algorithm to eliminate traits with redundant information content using the asymmetric measure of association Goodman and Kruskal's λ. We reduced the number of traits from 125 to 59 without any information loss. For the classification, we applied a random forest (RF) method. In this way, 93% of the cultivars were correctly classified using only four traits of the data set. To our knowledge, only a few studies applied a trait elimination algorithm similar to ours in ampelography that can be used for other biological data sets of similar structure. The classification results give a morphological explanation to several cultivars from the Carpathian Basin, a territory where all three Vitis vinifera L. geographical groups, occidentalis, orientalis and pontica, are represented. We found that the information-loss-avoiding data reduction method we applied in our study solved the redundancy-caused interdependencies and provided a suitable dataset for classifying grapevine genotypes. For example, this method may successfully be applied in digital image analysis-based traditional morphometric investigations in ampelography.

The genome sequence of the apple, Malus domestica (Suckow) Borkh., 1803

We present genome assemblies from four Malus domestica cultivars (the apple; Streptophyta; Magnoliopsida; Rosales; Rosaceae). The genome sequences are 643–653 megabases in span. The greater part of each assembly length (99.24–99.74%) is scaffolded into 17 chromosomal pseudomolecules. The mitochondrial and plastid genomes were also assembled and are 400 kilobases and 167 kilobases in length respectively.

microRNA172 targets APETALA2 to regulate flavonoid biosynthesis in apple (Malus domestica)

MicroRNA172 (miR172) plays a role in regulating a diverse range of plant developmental processes, including flowering, fruit development and nodulation. However, its role in regulating flavonoid biosynthesis is unclear. In this study, we show that transgenic apple plants over-expressing miR172 show a reduction in red coloration and anthocyanin accumulation in various tissue types. This reduction was consistent with decreased expression of APETALA2 homolog MdAP2_1a (a miR172 target gene), MdMYB10, and targets of MdMYB10, as demonstrated by both RNA-seq and qRT-PCR analyses. The positive role of MdAP2_1a in regulating anthocyanin biosynthesis was supported by the enhanced petal anthocyanin accumulation in transgenic tobacco plants overexpressing MdAP2_1a, and by the reduction in anthocyanin accumulation in apple and cherry fruits transfected with an MdAP2_1a virus-induced-gene-silencing construct. We demonstrated that MdAP2_1a could bind directly to the promoter and protein sequences of MdMYB10 in yeast and tobacco, and enhance MdMYB10 promotor activity. In Arabidopsis, over-expression of miR172 reduced flavonoid (including anthocyanins and flavonols) concentration and RNA transcript abundance of flavonoid genes in plantlets cultured on medium containing 7% sucrose. The anthocyanin content and RNA abundance of anthocyanin genes could be partially restored by using a synonymous mutant of MdAP2_1a, which had lost the miR172 target sequences at mRNA level, but not restored by using a WT MdAP2_1a. These results indicate that miR172 inhibits flavonoid biosynthesis through suppressing the expression of an AP2 transcription factor that positively regulates MdMYB10.

1-Methylcyclopropene maintains postharvest quality in Norwegian apple fruit

Norwegian fruit production is mostly destined for the local market and can suffer from poor-quality retention during storage. 1-Methylcyclopropene (1-MCP) is an inhibitor of ethylene perception used to maintain the physical and functional quality of pome fruit. Extensive work has been carried out on the effect of 1-MCP on apples, but not on cultivars grown in Norway. In this work, the potential of 1-MCP application (0.625 ml l ^-1 for 24 h at 0 ± 1℃) for ripening control of the apple cultivars 'Aroma', 'Red Gravenstein', and 'Summered' was studied during 1 and 1.5 months of cold storage; both scenarios were followed by five days of shelf life. The application of 1-MCP reduced softening by an average of 12% in 'Aroma', 'Red Gravenstein', and 'Summered' apples when cold stored for both 1 and 1.5 months as compared to control. External colour remained similar to initial values in 1-MCP fruit when compared to control apples, which presented a significant skin darkening. This indicated a delay in the ripening process. 1-MCP treatment did not affect total soluble solids content. 'Aroma' samples treated with 1-MCP showed a low sucrose hydrolysis, indicating a slower ripening process. This work confirms that 1-MCP postharvest treatment shows great potential for maintenance of apple cvs. in Norway during cold storage and shelf life.

Comparison of Common Analytical Methods for the Quantification of Total Polyphenols and Flavanols in Fruit Juices and Ciders

Multiple analytical methods are used for quantification of total polyphenols and total flavanols in fruit juices and beverages. Four methods were evaluated in this study: Folin-Ciocalteu (F-C), Lowenthal permanganate (L-P), 4-dimethylaminocinnamaldehyde (DMAC), and the bovine serum albumin (BSA) precipitation method. Method validation parameters, including working range, limit of detection, limit of quantitation, precision (repeatability), accuracy, and specificity, were assessed and compared. The F-C method was not specific to polyphenols, and the L-P method had the widest working range but lacked accuracy. The DMAC method was the most specific to flavanols, and the BSA method was not suitable for quantification of smaller flavanols, such as catechin and epicatechin. Quantitative performance was evaluated using commercial fruit juice samples (n = 14), apple juice samples of different cultivars (n = 22), and commercial ciders (n = 17). In general, the L-P titration method and DMAC method resulted in higher quantitative values than the F-C method and BSA precipitation method, respectively. However, ratios of results obtained by the L-P and F-C method ranged from 1 to 28, and ratios of results obtained by the DMAC and BSA precipitation method ranged from <1 to 280. This tremendous variation is likely due to variation in polyphenol composition and sample matrix. This information provides perspective for comparison of results obtained through these different methods, and a basis for choosing the most appropriate analytical method for quantification of polyphenols to address a specific research question when working with commercial fruit juice, apple juice from different apple cultivars, and commercial ciders. PRACTICAL APPLICATION: This study compared results obtained when four common polyphenol quantification methods were applied to a diverse selection of fruit juices and beverages with distinct polyphenol composition and sample matrix. The matrix and polyphenol composition of the samples significantly influenced the results. Our findings can help manufacturers of fruit-based products choose the most appropriate analytical method for polyphenol quantification as part of a quality assurance program or to convey information on dietary polyphenol content to consumers. An assessment of analytical method validation parameters is provided for each of the four methods, which will help users of these methods to understand their limitations.

Ancient Danish Apple Cultivars-A Comprehensive Metabolite and Sensory Profiling of Apple Juices

In recent decades, intensive selective breeding programs have allowed the development of disease-resistant and flavorsome apple cultivars while leading to a gradual decline of a large number of ancient varieties in many countries. However, the re-evaluation of such cultivars could lead to the production new apple-based products with health beneficial properties and/or unique flavor qualities. Herein, we report the comprehensive characterization of juices obtained from 86 old, mostly Danish, apple cultivars, by employing traditional analysis (ion chromatography, °Brix, headspace gas chromatography/mass spectrometry (GC-MS), and panel test evaluation) as well as an innovative nuclear magnetic resonance (NMR)-based screening method developed by Bruker for fruit juices, known as Spin Generated Fingerprint (SGF) Profiling™. Principal component analysis showed large differences in aroma components and sensory characteristics, including odd peculiar odors and flavors such as apricot and peach, and very different levels of phenolic compounds, acids and sugars among the analyzed juices. Moreover, we observed a tendency for late-season juices to be characterized by higher °Brix values, sugar content and they were perceived to be sweeter and more flavor intense than early-season juices. Our findings are useful for the production of specialty vintage-cultivar apple juices or mixed juices to obtain final products that are characterized both by healthy properties and peculiar sensory attributes.

Juice Clarification with Pectinase Reduces Yeast Assimilable Nitrogen in Apple Juice without Affecting the Polyphenol Composition in Cider

The term "cider" refers to the fermented beverage produced from apples. The rapid growth in the cider industry coupled with the prominence of traditional, or craft, approaches emphasizes the need for research on cider production. A common problem in ciders is the production of sulfur off-aromas by yeast during fermentation. Prefermentation juice clarification has the potential to reduce the occurrence of unwanted sulfur off-aromas. Concerns that prefermentation juice clarification will reduce the yeast assimilable nitrogen (YAN) and polyphenols in the juice have limited the application of this practice by cider makers. In this study, 3 clarification methods were applied to 'York' apple juice, that is, static settling, centrifugation, and pectinase. Raw (control) and clarified juice were fermented to cider, and the impact of clarification on the physicochemical parameters, amino acids and polyphenol content of the juice and cider was assessed. Juice clarification by pectinase decreased YAN by 50%, while static settling and centrifugation increased the concentration of most amino acids by 83%. All clarification treatments lowered the concentration of total polyphenols in the juice (from 60% to 30%, P < 0.05) and affected the individual polyphenols in the juice but these changes were not evident in the ciders. These findings demonstrate that prefermentation juice clarification results in changes in the chemistry profiles of apple juice. These changes were however not evident in the ciders. This approach therefore has the potential to limit the production of sulfur off-aromas during cider production without adverse effects on quality. PRACTICAL APPLICATION: Clarification of apple juice changes polyphenol and nitrogen contents, factors known to affect cider quality. However, these differences in the chemical profile of apple juice may not be reflected in the finished ciders. These findings demonstrate that juice clarification can be employed in cider making without affecting the quality. Cider makers should not assume that changes in apple juice imparted by clarification will be reflected in the finished ciders. Outcomes should be measured in finished ciders, rather than juice to accurately evaluate effects of the processing steps on quality.