Fermentation Monitoring Research Articles

Validation of an Optical Technology for the Determination of pH in Milk during Yogurt Manufacture.

Current systems that allow inline pH control in the fermented dairy industry have drawbacks, such as protein adhesion on the non-glass pH probes, measurement distortion, frequent recalibration needs, and sensitivity to extreme pH conditions encountered during clean-in-place operations. Therefore, the objective of this study was to validate the feasibility of estimating the pH of milk during the yogurt making process by using a NIR light backscatter sensor measuring under different fermentation temperatures and milk protein concentrations using a mathematical model that correlates the light scatter signal with pH. Three replications of the experiment with two protein concentrations (3.5 and 4.0%) and two fermentation temperatures (43 and 46 °C) were used to validate this inline pH prediction model. Continuous and discontinuous measurements of pH were collected as a reference during fermentation, simultaneously with the light backscatter data acquisition. Also, the effect of adjusting the initial voltage gain of the light scatter device on the accuracy of the pH prediction model was evaluated. Temperature and initial voltage were the main factors affecting the fitting accuracy of the model. The adjustment of the initial voltage gain improved the pH prediction model fit. The model has been successfully validated for both continuous and discontinuous measurements of pH, with SEP values < 0.09 pH units and CV < 1.78%. The proposed optical inline and non-destructive method was feasible for inline pH monitoring of milk fermentation, avoiding traditional manual pH measurement.

Comparison of Different Spectral Ranges to Monitor Alcoholic and Acetic Fermentation of Red Grape Must Using FT-NIR Spectroscopy and PLS Regression

AbstractWine vinegar is produced through a two-phase fermentation of grape must: initially, yeast converts grape sugars into ethanol, and subsequently, acetobacteria oxidize ethanol into acetic acid. This process, spanning weeks when conducted by surface fermentation, requires constant monitoring of ethanol and total acidity levels. To enhance the quality and efficiency of process monitoring, vinegar production is shifting to faster, environmentally sustainable methods. Near-infrared (NIR) spectroscopy, recognized for its non-invasiveness and speed, is ideal for online implementation in process control. This study tracked dual fermentation in red grape must over an extended period, monitoring two different batches simultaneously to assess fermentation kinetics and reproducibility. Ethanol content and total acidity were analyzed in fermenting musts throughout the whole fermentation process using both classical laboratory analyses and FT-NIR spectroscopy. Principal Component Analysis (PCA) was used to explore the spectral dataset, then Partial Least Squares (PLS) was used to develop calibration models for predicting ethanol and acidity. The models calculated considering the entire spectral range were compared with those obtained for two narrower zones, where more cost-effective and easily miniaturizable sensors are available on the market. FT-NIR allowed to effectively determine ethanol content and acidity (R2Pred &gt; 0.98), both over the entire range (12,500–4000 cm−1, corresponding to 800–2500 nm) and in the 10,526–6060 cm−1 (950–1650 nm) region. Although less satisfactory, still acceptable results were obtained in the 12,500–9346 cm−1 (800–1070 nm) region (R2Pred &gt; 0.81), confirming the potential for cost-effective devices in real-time fermentation monitoring.

Fermentation Analytical Technology (FAT): Monitoring industrial E. coli fermentations using absolute quantitative 1H NMR spectroscopy

BackgroundTo perform fast, reproducible, and absolute quantitative measurements in an automated manner has become of paramount importance when monitoring industrial processes, including fermentations. Due to its numerous advantages – including its inherent quantitative nature – Proton Nuclear Magnetic Resonance (1H NMR) spectroscopy provides an ideal tool for the time-resolved monitoring of fermentations. However, analytical conditions, including non-automated sample preparation and long relaxation times (T1) of some metabolites, can significantly lengthen the experimental time and make implementation in an industrial set up unfeasible. ResultsWe present a high throughput method based on Standard Operating Procedures (SOPs) and 1H NMR, which lays the foundation for what we call Fermentation Analytical Technology (FAT). Our method was developed for the accurate absolute quantification of metabolites produced during Escherichia coli industrial fermentations. The method includes: (1) a stopped flow system for non-invasive sample collection followed by sample quenching, (2) automatic robot-assisted sample preparation, (3) fast 1H NMR measurements, (4) metabolites quantification using multivariate curve resolution (MCR), and (5) metabolites absolute quantitation using a novel correction factor (k) to compensate for the short recycle delay (D1) employed in the 1H NMR measurements. The quantification performance was tested using two sample types: buffer solutions of chemical standards and real fermentation samples. Five metabolites – glucose, acetate, alanine, phenylalanine and betaine – were quantified. Absolute quantitation ranged between 0.64 and 3.40 mM in pure buffer, and 0.71–7.76 mM in real samples. SignificanceThe proposed method is generic and can be straight forward implemented to other types of fermentations, such as lactic acid, ethanol and acetic acid fermentations. It provides a high throughput automated solution for monitoring fermentation processes and for quality control through absolute quantification of key metabolites in fermentation broth. It can be easily implemented in an at-line industrial setting, facilitating the optimization of the manufacturing process towards higher yields and more efficient and sustainable use of resources.

Optics miniaturization strategy for demanding Raman spectroscopy applications

Raman spectroscopy provides non-destructive, label-free quantitative studies of chemical compositions at the microscale as used on NASA’s Perseverance rover on Mars. Such capabilities come at the cost of high requirements for instrumentation. Here we present a centimeter-scale miniaturization of a Raman spectrometer using cheap non-stabilized laser diodes, densely packed optics, and non-cooled small sensors. The performance is comparable with expensive bulky research-grade Raman systems. It has excellent sensitivity, low power consumption, perfect wavenumber, intensity calibration, and 7 cm−1 resolution within the 400–4000 cm−1 range using a built-in reference. High performance and versatility are demonstrated in use cases including quantification of methanol in beverages, in-vivo Raman measurements of human skin, fermentation monitoring, chemical Raman mapping at sub-micrometer resolution, quantitative SERS mapping of the anti-cancer drug methotrexate and in-vitro bacteria identification. We foresee that the miniaturization will allow realization of super-compact Raman spectrometers for integration in smartphones and medical devices, democratizing Raman technology.

Forecasting wine phenolic composition from infrared spectra of grapes extracts and monitoring of fermentations with optimised time-specific prediction models

Incorporating monitoring and forecasting technologies has the potential to enhance the different aspects of winemaking. By monitoring certain chemical parameters, it is possible to evaluate the progress of the fermentation and phenolic extraction, while also detecting deviations from expected behaviour. This study aims to investigate the development of a rapid forecasting method and different strategies to improve the accuracy of process monitoring. The first part involves the introduction of partial least squares (PLS) calibration models able to forecast the trajectories of phenolic extraction during fermentation using grape extract spectra. Additionally, novel methods for reoptimizing calibration models are proposed, with the intention of increasing the prediction accuracy. Whilst this method is transferable to PLS models, principal component regression (PCR) models were used instead as a more simplistic way to test the new methods. First, dynamic PCR linear regression coefficients were adapted over time to capture the compound evolution in the wine matrix. Secondly, PCR calibration models were reoptimized using interpolated spectral and reference data. The forecasting PLS calibration models have shown promising results for total anthocyanin content (mg/L), colour density, polymeric pigments (mg/L), MCP tannins (mg/L), and total phenolic index with overall R2 values on the validation set above 0.6, and low RSMEV values. Using interpolated data to re-optimise models was found to be more effective than global models or those using the calculated regression coefficients methods for predicting the concentration of a particular compound at a certain time point. While R2 values are suboptimal for all considered PCR models, in 72% of the cases analysed, the interpolation method for re-optimisation showed RMSEV values between 6% and 83% lower than those reported for a more traditional PCR model and other re-optimisation strategies. Interpolated models have shown consistency in predicting values, along with good performance metrics, indicating that they could be applied to samples for which calibration models do not exist.

Machine learning algorithms for in-line monitoring during yeast fermentations based on Raman spectroscopy

Given the intricacies and nonlinearity inherent to industrial fermentation systems, the application of process analytical technology presents considerable benefits for the direct, real-time monitoring, control, and assessment of synthetic processes. In this study, we introduce an in-line monitoring approach utilizing Raman spectroscopy for ethanol production by Saccharomyces cerevisiae. Initially, we employed feature selection techniques from the realm of machine learning to reduce the dimensionality of the Raman spectral data. Our findings reveal that feature selection results in a noteworthy reduction of over 90% in model training time, concurrently enhancing the predictive performance of glycerol and cell concentration by 14.20% and 17.10% at the root mean square error (RMSE) level. Subsequently, we conducted model retraining using 15 machine learning algorithms, with hyperparameters optimized through grid search. Our results illustrate that the post-hyperparameter adjustment model exhibits improvements in RMSE for ethanol, glycerol, glucose, and biomass by 9.73%, 4.33%, 22.22%, and 13.79%, respectively. Finally, specific machine learning algorithms, namely BaggingRegressor, Support Vector Regression, BayesianRidge, and VotingRegressor, were identified as suitable models for predicting glucose, ethanol, glycerol, and cell concentrations, respectively. Notably, the coefficient of determination (R2) ranged from 0.89 to 0.97, and RMSE values ranged from 0.06 to 2.59 g/L on the testing datasets. The study highlights machine learning's effectiveness in Raman spectroscopy data analysis for improved industrial fermentation monitoring, enhancing efficiency, and offering novel modeling insights.

Optimizing the winemaking process: NIR spectroscopy and e-nose analysis for the online monitoring of fermentation.

In the winemaking process, the rapid determination of specific quality parameters such as sugar content, pH, acidity, concentrations of phenolic compounds, anthocyanins and volatile organic compounds is crucial for high-quality wine production. Traditional analytical methods allow for precise quantification of these parameters but are time-consuming and expensive. This article explores the potential application of non-destructive analytical technique (NDAT) (near infra-red [NIR] and e-nose), as efficient alternatives for online monitoring of fermentation working on two different winemaking tanks and applying chemometrics to develop predictive models to correlate non-destructive and analytical data. NIR measurements have been used to build principal components regression models, showing good prediction capability for polyphenols, anthocyanins, glucose and fructose. Both offline and online e-nose applications demonstrate good capability of discriminating different fermentation phases, in agreement with aromatic profile changes observed via gas chromatography-mass spectrometry analysis. Moreover, correlation analysis reveals the potential of quartz microbalances, Taguchi Gas Sensors and H2 S sensors in predicting the concentration of compounds of great interest for winemaking (e.g. C6 alcohols, ketones, terpenes and ethyl esters) highlighting the robust connection between sensor data and specific chemical classes. This research aims to showcase the potential employment of NDAT for online monitoring the evolution of must composition during fermentation. The proposed methods could potentially fulfil a longstanding requirement of winemakers, enabling them to closely monitor fermentation allowing the timely making of important technical decisions aimed at achieving oenological objectives in wine production. © 2024 Society of Chemical Industry.

New Insights into the Comprehensive System of Thermodynamic Sensors and Electronic Nose and Its Practical Applications in Dough Fermentation Monitoring.

This study focuses on an applicability of the device designed for monitoring dough fermentation. The device combines a complex system of thermodynamic sensors (TDSs) with an electronic nose (E-nose). The device's behavior was tested in experiments with dough samples. The configuration of the sensors in the thermodynamic system was explored and their response to various positions of the heat source was investigated. When the distance of the heat source and its intensity from two thermodynamic sensors changes, the output signal of the thermodynamic system changes as well. Thus, as the distance of the heat source decreases or the intensity increases, there is a higher change in the output signal of the system. The linear trend of this change reaches an R2 value of 0.936. Characteristics of the doughs prepared from traditional and non-traditional flours were successfully detected using the electronic nose. To validate findings, the results of the measurements were compared with signals from the rheofermentometer Rheo F4, and the correlation between the output signals was closely monitored. The data after statistical evaluation show that the measurements using thermodynamic sensors and electronic nose directly correlate the most with the measured values of the fermenting dough volume. Pearson's correlation coefficient for TDSs and rheofermentometer reaches up to 0.932. The E-nose signals also correlate well with dough volume development, up to 0.973. The data and their analysis provided by this study declare that the used system configuration and methods are fully usable for this type of food analysis and also could be usable in other types of food based on the controlled fermentation. The system configuration, based on the result, will be also used in future studies.

Bread fermentation monitoring through NIR spectroscopy and PLS-DA. Determining the optimal fermentation point in bread doughs

Bread fermentation monitoring through NIR spectroscopy and PLS-DA. Determining the optimal fermentation point in bread doughs

Bacterial community structure, predicted metabolic activities, and formation of volatile compounds attributed to Malaysian fish sauce flavour.

The fermentation of Malaysian fish sauce (budu) varies from one to twelve months depending on the producer, resulting in inconsistent quality. The microbiota, their predicted metabolic pathways and volatile metabolites profiles were determined at different stages of budu fermentation. Budu fermented for 1 and 3months were characterized by the presence of Gram negative Enterobacterales, Gammaproteobacteria, and Fusobacteriaceae, which continuously decrease in abundance over fermentation time. The metabolic pathways prediction grouped 1- and 3- month budu in a cluster enriched with degradation reactions. 6-month budu were dominated by Halanaerobium and Staphylococcus, while the 12-month were dominated by Lentibacillus, Bacilli, and Halomonas. Biosynthesis-type predicted pathways involving protein and lipid derivatives were enriched in 6- and 12-month fermented budu, accumulating 2,6-dimethylpyrazine, methyl 2-ethyldecanoate, 2-phenylacetaldehyde, 3-methylbutanal, and 3-methylbutanoic acid. These compounds may indicate budu maturity and quality. This result may assist as a reference for quality control and fermentation monitoring.

Monitoring of simultaneous saccharification and fermentation of ethanol by multi-source data deep fusion strategy based on near-infrared spectra and electronic nose signals

Fuel ethanol represents a future energy trajectory, and the simultaneous saccharification and fermentation (SSF) technique emerges as the principal approach for ethanol production. This scholarly inquiry offers an innovative means to monitor the SSF process for ethanol meticulously. Employing a profound fusion strategy that effectively amalgamates diverse data sources. The convolutional neural network and recurrent neural network (RNN) architectures are thoughtfully crafted and designed to enable autonomous feature self-learning from near-infrared spectra and electronic nose data. These intricately devised networks further implement data fusion strategies at the granular level of features. Ultimately, a deep fusion correction model was devised and rigorously validated using two distinct data sources, namely near-infrared spectroscopy and electronic nose data. The obtained results demonstrate a discernible improvement in the overall predictive accuracy of the model when employing the fusion feature strategy, surpassing the model constructed solely on a single technical data source. Regarding the monitoring of ethanol content, the optimal RNN fusion model exhibited remarkable performance metrics, with a root mean square error of prediction (RMSEP) value of 3.2265, a coefficient of determination (R2) value of 0.9880, and a relative percent deviation (RPD) value of 9.2662. In terms of monitoring glucose content, the optimal RNN fusion model also demonstrated commendable performance, with the following respective parameters: RMSEP was 3.2770, R2 was 0.9840, and RPD was 8.0085. The overall results indicate that the multi-sensor data fusion strategy not only improves the performance of the model but also provides valuable insights into the fermentation process.

Utilization of the AIRMIXING M.I.™ System in Producing Red Wine without Added Sulphites

Sulphur dioxide (SO2) is usually used in winemaking due to its effectiveness as an antimicrobial and antioxidant agent. However, there is growing interest in finding alternatives to SO2 because of its adverse effects on human health. Therefore, in this work, a low-SO2-producing Saccharomyces cerevisiae strain was selected and the AIR-MIXINGTM M.ITM (A.M.) system, capable of determining a high extraction of polyphenols, was used to produce a red wine without added sulphites. A tank managed with the traditional pumping-over technique was used as a comparison. Microbiological and chemical monitoring of both fermentations performed by plate counts and HPLC analyses, respectively, did not indicate significant differences as regards the yeast growth kinetics and the degradation of the sugars, while it highlighted a faster extraction kinetics of colour and total polyphenols in the fermentation carried out with the A.M. system. Both experimental wines showed a total SO2 content &lt;10 mg/L, but in the wine produced with A.M., a higher content of the polymeric forms of anthocyanins and non-anthocyanin phenols was found in favour of a higher stability of the wine achieved in a shorter time than the control. Furthermore, a higher concentration of reduced glutathione, a compound well-known for its antioxidant activity, occurred in wine obtained with the A.M. system. In conclusion, the use of low-SO2-producing yeasts in combination with the A.M. system could be a suitable approach to produce wines without sulphites added.

In-fiber Mach-Zehnder interferometer as a promising tool for optical nose and odor prediction during the fermentation process.

In this paper, we present an in-fiber Mach-Zehnder interferometer (MZI) applied to coffee bean fermentation monitoring. Two MZIs, based on a combination of a fiber taper cascaded by a micro-tapered long-period fiber grating, were installed in a fermentation barrel to monitor the liquids and gases released during the fermentation process. During this process, a variety of odors arise due to the yeast activity and their classification is important to decide when to stop the fermentation process. In this work, we show that the in-fiber MZIs are good candidates for optical noses in this scenario.

γ-polyglutamic acid fermentation monitoring with ATR-FTIR spectroscopy based on a shallow convolutional neural network combined with data augmentation and attention module

The potential of attenuated total reflection Fourier transform infrared (ATR-FTIR) spectroscopy to online real-time monitor the γ-polyglutamic acid (γ-PGA) fermentation process by measuring the concentration of two key components, glucose and sodium glutamate, was investigated. Partial least squares regression (PLSR) and convolutional neural network (CNN) were selected as the multivariate calibration model to predict fermentation parameters. To solve the small sample training problem with CNN, the data augmentation strategy was firstly combined with a shallow CNN to generate a new CNN model termed as DA-CNN. To improve the performance of DA-CNN further, squeeze-and-excitation (SE) module of the attention mechanism was integrated into DA-CNN to form a DA-CNN variant termed as DA-SE-CNN, which could capture the different importance of extracted channel-wise feature to improve feature representation ability. Experiments were conducted on the γ-PGA dataset, and the results showed that the nonlinear models (CNN, DA-CNN and DA-SE-CNN) performed better than the linear method (PLSR) in all cases except the case of sodium glutamate prediction with CNN. Additionally, both DA-CNN and DA-SE-CNN perform better than CNN in all cases with limited training data. Especially DA-SE-CNN gave the best result with excellent prediction accuracy, which indicated that ATR-FTIR combined with nonlinear regression tool (DA-SE-CNN) as a rapid method to monitor γ-PGA fermentation process is feasible.

Acidogenic Fermentation of Brewers’ Spent Grain Monitored through Two-Dimensional Fluorescence Spectroscopy

Biological systems are commonly controlled and monitored through offline and time-consuming tools, which often impairs an effective and real-time response to counteract system disturbances. The feasibility of using two-dimensional (2D) fluorescence spectroscopy as a non-invasive, non-destructive, and real-time procedure to monitor the acidogenic fermentation of brewer’s spent grain (BSG) in a granular sludge reactor was evaluated. For that, the effect of pH fluctuations on the system response was used as a model to ascertain the 2D fluorescence spectroscopy applicability to monitor the process performance, namely, to predict the fermentation products (FP) and the soluble protein (SProt) concentrations in the effluent stream through mathematical analysis. The pH fluctuations over the course of the reactor’s operation altered the granules’ microbiome composition, leading to different effluent FP profiles. Fluorescence excitation–emission matrices (EEMs) were used with projection to latent structures (PLS) modeling to predict the FP and SProt concentrations in the effluent with average errors below 0.75 and 0.43 g L–1, respectively. Both models were able to capture the tendency of the data even when the accuracy of prediction was not so high. The combined approach of using 2D fluorescence spectroscopy and mathematical analysis seemed promising for real-time monitoring of the acidogenic fermentation of complex substrates.

Machine learning and domain adaptation to monitor yoghurt fermentation using ultrasonic measurements

In manufacturing environments, real-time monitoring of yoghurt fermentation is required to maintain an optimal production schedule, ensure product quality, and prevent the growth of pathogenic bacteria. Ultrasonic sensors combined with machine learning models offer the potential for non-invasive process monitoring. However, methods are required to ensure the models are robust to changing ultrasonic measurement distributions as a result of changing process conditions. As it is unknown when these changes in distribution will occur, domain adaptation methods are needed that can be applied to newly acquired data in real-time. In this work, yoghurt fermentation processes are monitored using non-invasive ultrasonic sensors. Furthermore, a transmission based method is compared to an industrially-relevant non-transmission method which does not require the sound wave to travel through the fermenting yoghurt. Three machine learning algorithms were investigated including fully-connected neural networks, fully-connected neural networks with long short-term memory layers, and convolutional neural networks with long short-term memory layers. Three real-time domain adaptation strategies were also evaluated, namely; feature alignment, prediction alignment, and feature removal. The most accurate method (mean squared error of 0.008 to predict pH during fermentation) was non-transmission based and used convolutional neural networks with long short-term memory layers, and a combination of all three domain adaption methods.

Microsatellite typing of Lachancea thermotolerans for wine fermentation monitoring

Climate change is causing a lack of acidity during winemaking and oenologists use several solutions to cope with such a problem. Lachancea thermotolerans, which has the potential to tolerate the harsh physicochemical conditions of wine, has emerged as a promising alternative for pH management during winemaking and, currently, it is the most valuable yeast used for acidity control in wine. In this work a manageable method for L. thermotolerans genotyping based on a multiplexed microsatellite amplification in 6 different loci was developed. The proposed method was used to distinguish between 103 collection strains obtained from different geographical and isolation sources, and then challenged against a 429 L. thermotolerans isolates from several wineries and harvests. The procedure was also tested for fermentation monitoring and strain implantation. This approach was conceived to simplify the methodology available for L. thermotolerans genotyping, making it easy for applying in wine-related laboratories. This method can be applied to distinguish between L. thermotolerans strains in selection programs and to follow implantation of inoculated strains during winemaking with optimal results.

Arduino Soft Sensor for Monitoring Schizochytrium sp. Fermentation, a Proof of Concept for the Industrial Application of Genome-Scale Metabolic Models in the Context of Pharma 4.0

Schizochytrium sp. is a microorganism cultured for producing docosahexaenoic acid (DHA). Genome-scale metabolic modeling (GEM) is a promising technique for describing gen-protein-reactions in cells, but with still limited industrial application due to its complexity and high computation requirements. In this work, we simplified GEM results regarding the relationship between the specific oxygen uptake rate (−rO2), the specific growth rate (µ), and the rate of lipid synthesis (rL) using an evolutionary algorithm for developing a model that can be used by a soft sensor for fermentation monitoring. The soft sensor estimated the concentration of active biomass (X), glutamate (N), lipids (L), and DHA in a Schizochytrium sp. fermentation using the dissolved oxygen tension (DO) and the oxygen mass transfer coefficient (kLa) as online input variables. The soft sensor model described the biomass concentration response of four reported experiments characterized by different kLa values. The average range normalized root-mean-square error for X, N, L, and DHA were equal to 1.1, 1.3, 1.1, and 3.2%, respectively, suggesting an acceptable generalization capacity. The feasibility of implementing the soft sensor over a low-cost electronic board was successfully tested using an Arduino UNO, showing a novel path for applying GEM-based soft sensors in the context of Pharma 4.0.

Advantages of label free method in comparison with 2DE proteomic analysis of Butyrivibrio fibrisolvens 3071 grown on different carbon sources

The aim of this study was the comparison of label-free method with 2DE to other analytical method of bacterium Butyrivibrio fibrisolvens extracellular protein samples. Label-free quantification (LF) method performed in this work was compared with previously obtained results of proteomic analysis using two-dimensional electrophoresis (2DE) and nano-liquid chromatography coupled with mass spectroscopy (nLC/MS). B. fibrisolvens as an important plant fibre degrader was cultivated on four different carbon sources (including xylan, xylose, glucose, and a mixture of xylan with glucose). The impact of growth substrate on protein profile was assessed by six pair-wise comparisons evaluating the significantly differently abundant extracellular proteins. Gel-free and gel-based methods resulted in substantially dissimilar results. The LC-MS/MS approach detected substrate-dependent differences in transport and binding membrane proteins (TBP) and nucleotidase, while the 2DE approach detected substrate effect on proteins included in protein synthesis and butyrate synthesis. On the other hand, both methods observed differentially regulated proteins involved in the glycolytic pathway, however, the only shared enzyme (protein detected both by 2DE and LC-MS/MS approach) was fructose-bisphosphate aldolase. The LC-MS/MS approach detected a high abundance of separated peptides. However, it cannot be easily considered as supreme to 2DE analysis. Both methods differ in sample preparation, their advantages and limitations. The study findings indicate these methods can complement each other and together they elucidate better the metabolic functions of B. fibrisolvens. HIGHLIGHTS Comparison of LF and 2DE analysis showed limitations connected with a narrow optimised pH interval and number of affected proteins in 2DE in contrary to LF results. Butyrivibrio fibrisolvens culture showed the same proteomic changes in the presence of different saccharides as the whole rumen fluid in vivo. Selected proteins could be used for the monitoring of polysaccharides fermentation in the rumen and be used for optimisation hemicellulose degradation and overall feed utilisation.

Visualizing chemical indicators: Spatial and temporal quality formation and distribution during black tea fermentation

Fermentation is a key black tea processing step and makes an important contribution to quality formation. Current approaches to fermentation monitoring are costly or laboratory-based. Here, we first evaluated the potential of at-line computer vision for detecting fermentation quality in a tea factory. A self-built industrial camera was used to collect tea samples at various fermentation durations. The correlations of color variables that were extracted from the images with key quality indicators in the tea samples were verified. Subsequently, partial least-squares regression models based on the color variables showed high prediction accuracy with residual prediction deviation values of 4.13, 3.53, and 3.39 for catechins, theaflavins and chlorophylls, respectively. Finally, the spatial and temporal distributions of indicators during fermentation were mapped to visualize the fermentation quality. This study realized low-cost, at-line and real-time detection for black tea fermentation, which provides technical support for the industrial and intelligent production of black tea.