Application of bio-inspired optimization algorithms in food processing

A hybrid RSM-BPNN-GA approach for optimizing ultrasound-assisted deep eutectic solvents extraction conditions for Mesona chinensis benth. and investigation of the extraction mechanism

Mesona chinensis Benth (MCB) is the source of the most commonly consumed herbal beverage in Southeast Asia and China and is thus an economically important agricultural plant. Therefore, optimal extraction and production procedures have significant commercial value. Currently, in terms of green chemistry, researchers are investigating the use of greener solvents and innovative extraction techniques to increase extract yields. This study represents the first investigation of the optimal conditions for ultrasound-assisted deep eutectic solvent (DES) extraction from MCB. The major factors influencing ultrasound-assisted DESs were optimized using the response surface methodcentral-genetic algorithm-back propagation neural networks. This model demonstrated superior predictability and accuracy compared to the RSM model. Various types of DESs were used for the extraction of MCB constituents, with choline chloride-ethylene glycol resulting in the highest yield. The optimal conditions for maximal extraction were the use of choline chloride-ethylene glycol (1:4) as the solvent with a 40% water content, an extraction duration of 60 min at 60°C, and maintaining a leaf-to-solvent ratio of 20 mL/g. Noticeable enhancements in Van der Waals forces and more robust interactions between DESs and the target chemicals were observed relative to those seen with ethanol (70%, v/v) or water. This investigation not only introduced an environmentally friendly approach for highly efficient extraction from MCB but also identified the mechanisms underlying the improved extraction efficacy. These findings have the potential to contribute to the broader utilization of MCB and provide valuable insights into the extraction mechanisms utilizing deep eutectic solvents. PRACTICAL APPLICATION: This work describes an efficient and green ultrasound-assisted deep eutectic solvent (DES) method for Mesona chinensis Benth (MCB) extraction. Molecular dynamics was used to examine the intermolecular interactions between the solvent and the extracted compounds. It is anticipated that green and environmentally friendly solvents, such as DESs, will be used in further research on foods and their bioactive components. With the development of the herbal tea industry, new products made of MCB are becoming increasingly popular, thus gradually making it a research hotspot.

Achieving sustainability in heat drying processing: Leveraging artificial intelligence to maintain food quality and minimize carbon footprint

The food industry is a significant contributor to carbon emissions, impacting carbon footprint (CF), specifically during the heat drying process. Conventional heat drying processes need high energy and diminish the nutritional value and sensory quality of food. Therefore, this study aimed to investigate the integration of artificial intelligence (AI) in food processing to enhance quality and reduce CF, with a focus on heat drying, a high energy-consuming method, and offer a promising avenue for the industry to be consistent with sustainable development goals. Our finding shows that AI can maintain food quality, including nutritional and sensory properties of dried products. It determines the optimal drying temperature for improving energy efficiency, yield, and life cycle cost. In addition, dataset training is one of the key challenges in AI applications for food drying. AI needs a vast and high-quality dataset that directly impacts the performance and capabilities of AI models to optimize and automate food drying.

Disease Diagnosis Based on Improved Gray Wolf Optimization (IGWO) and Ensemble Classification

This paper introduces a simple strategy for diagnosing disease, which is called improved gray wolf optimization (IGWO) and ensemble classification. The proposed strategy consists of two sequential phases, which are; (i) Feature Selection Phase (FSP) and (ii) Ensemble Classification Phase (ECP). During the former, the most effective features for diagnosing disease are selected, while during the latter, the actual diagnosis takes place depending on voting of five different classifiers. The main contribution of this paper is a suggested modification for the traditional Gray Wolf Optimization (GWO), which is called Improved Gray Wolf Optimization (IGWO). As an optimization technique, the proposed IGWO is employed in the FSP for selecting the effective features. For evaluating, IGWO has been implemented using recent feature selection techniques as well as the proposed method. To accomplish the classification phase; ensemble classification has been used which uses several classification techniques such as; Naïve Bayes (NB), Support Vector Machines (SVM), Deep Neural Network (DNN), Decision Tree (DT), and K-Nearest Neighbors (KNN). Ensemble classification integrate several classifiers for improving prediction performance. Experimental results have shown that employing IGWO promotes the performance of the diagnosing strategy of different diseases in terms of precision, recall, and accuracy.

A Comprehensive Review of Bio-Inspired Optimization Algorithms Including Applications in Microelectronics and Nanophotonics

The application of artificial intelligence in everyday life is becoming all-pervasive and unavoidable. Within that vast field, a special place belongs to biomimetic/bio-inspired algorithms for multiparameter optimization, which find their use in a large number of areas. Novel methods and advances are being published at an accelerated pace. Because of that, in spite of the fact that there are a lot of surveys and reviews in the field, they quickly become dated. Thus, it is of importance to keep pace with the current developments. In this review, we first consider a possible classification of bio-inspired multiparameter optimization methods because papers dedicated to that area are relatively scarce and often contradictory. We proceed by describing in some detail some more prominent approaches, as well as those most recently published. Finally, we consider the use of biomimetic algorithms in two related wide fields, namely microelectronics (including circuit design optimization) and nanophotonics (including inverse design of structures such as photonic crystals, nanoplasmonic configurations and metamaterials). We attempted to keep this broad survey self-contained so it can be of use not only to scholars in the related fields, but also to all those interested in the latest developments in this attractive area.

Characterization of Probiotic Pichia sp. DU2-Derived Exopolysaccharide with Oil-in-Water Emulsifying and Anti-biofilm Activities

Probiotic-derived exopolysaccharides are considered as promising sources of carbohydrate with extensive applications in many industries. In the current study, yeast strains were isolated from chicken ingluvies and gizzard samples. According to molecular identification, EPS-producing yeast (Pichia sp. DU2) showed the most similarity to Pichia cactophila (99.67%). Pichia sp. DU2 showed probiotic properties. EPS of Pichia sp. DU2 showed emulsifying activity. The formed emulsions showed 53% (colza oil) and 100% (p-xylene) stability after 24 h. These emulsions were oil-in-water and have stability in the presence of NaCl, KCl, and also acidic and basic conditions. Also, the EPS showed anti-biofilm (29.7-47.6% and 19.06-55.26% against B. cereus and Y. enterocolitica, respectively) and flocculating activities (31.4%). FT-IR showed the presence of various functional groups in EPS structure. Also, its heteropolysaccharide nature was revealed in ¹H-NMR and HPLC analysis. This emulsifying EPS showed significant thermal stability and negative zeta potential, which make it a promising carbohydrate for various industries. Finally, according to the predicted model, the maximal EPS production was achieved at reaction time 36 h, pH 6, yeast extract concentration 1.0%, and sucrose concentration 5%. Pichia sp. DU2 with probiotic properties and producing EPS with emulsifying, anti-biofilm, and flocculating activities can be considered as promising yeast strain in various industries like food and pharmaceutical industries.

Modeling and optimizing in vitro percentage and speed callus induction of carrot via Multilayer Perceptron-Single point discrete GA and radial basis function

Background

Callus induction is the first step in optimizing plant regeneration. Fit embryogenesis and shooting rely on callus induction. In addition, using artificial intelligence models in combination with an algorithm can be helpful in the optimization of in vitro culture. The present study aimed to evaluate the percentage and speed of callus induction optimization in carrot with a Multilayer Perceptron-Single point discrete genetic algorithm (GA).

Materials and methods

In this study, the outputs included callus induction percentage and speed, while inputs were different types and concentrations of plant growth regulator (0. 5, 0.2 mg/l 2,4-D, 0.3, 0.2, 0.5 mg/l BAP, 1, 0.2 mg/l Kin, and 2 mg/l NAA), different explants (shoot, root, leaf, and nodal), a different concentration compound of MS medium (1 × MS, 4× MS, and 8× MS) and time of sampling. The data were obtained in the laboratory, and multilayer perceptron (MLP) and radial basis function (RBF), two well-known ANNs, were employed to model. Then, GA was used for optimization, and sensitivity analysis was performed to indicate the inputs’ importance.

Results

The results showed that MLP had better prediction efficiency than RBF. Based on the results, R2 in training and testing data was 95 and 95% for the percentage of callus induction, while it was 94 and 95% for the speed of callus induction, respectively. In addition, a concentration compound of MS had high sensitivity, while times of sampling had low sensitivity. Based on the MLP-Single point discrete GA, the best results were obtained for shoot explants, 1× MS media, and 0.5 mg/l 2, 4-D + 0.5 mg/l BAP. Further, a non-significant difference was observed between the test result and predicted MLP.

Conclusions

Generally, MLP-Single point discrete GA is considered a potent tool for predicting treatment and fit model results used in plant tissue culture and selecting the best medium for callus induction.

Anti-biofilm efficacy of green-synthesized ZnO nanoparticles on oral biofilm: In vitro and in silico study

The development of biofilm on the biotic and abiotic surfaces is the greatest challenge for health care sectors. At present times, oral infection is a common concern among people with an unhealthy lifestyle and most of these biofilms-associated infections are resistant to antibiotics. This has increased a search for the development of alternate therapeutics for eradicating biofilm-associated infection. Nanobiotechnology being an effective way to combat such oral infections may encourage the use of herbal compounds, such as bio-reducing and capping agents. Green-synthesis of ZnO nanoparticles (ZnO NP) by the use of the floral extract of Clitoria ternatea, a traditionally used medicinal plant, showed stability for a longer period of time. The NPs as depicted by the TEM image with a size of 10 nm showed excitation spectra at 360 nm and were found to remain stable for a considerable period of time. It was observed that the NPs were effective in the eradication of the oral biofilm formed by the major tooth attacking bacterial strains namely Porphyromonsas gingivalis and Alcaligenes faecalis, by bringing a considerable reduction in the extracellular polymeric substances (EPS). It was observed that the viability of the Porphyromonsas gingivalis and Alcaligenes faecalis was reduced by NP treatment to 87.89 ± 0.25% in comparison to that of amoxicillin. The results went in agreement with the findings of modeling performed by the use of response surface methodology (RSM) and artificial neural network (ANN). The microscopic studies and FT-IR analysis revealed that there was a considerable reduction in the biofilm after NP treatment. The in silico studies further confirmed that the ZnO NPs showed considerable interactions with the biofilm-forming proteins. Hence, this study showed that ZnO NPs derived from Clitoria ternatea can be used as an effective alternative therapeutic for the treatment of biofilm associated oral infection.

Ultrasonic Extraction Process of Polysaccharides from Dendrobium nobile Lindl.: Optimization, Physicochemical Properties and Anti-Inflammatory Activity

To optimize the ultrasonic extraction process of polysaccharides from Dendrobium nobile Lindl. (DNP), the extraction method was conducted through a single-factor test and the response-surface methodology (RSM). With the optimal extraction process (liquid–solid ratio of 40 mL/g, ultrasonic time of 30 min, and ultrasonic power of 400 W), the maximum extraction yield was 5.16 ± 0.41%. DNP1 and DNP2 were then fractionated via DEAE-QFF and Sephacryl S-300 HR chromatography. The molecular weight (Mw) of DNP1 was identified as 67.72 kDa, composed of Man (75.86 ± 0.05%) and Glc (24.14 ± 0.05%), and the Mw of DNP2 was 37.45 kDa, composed of Man (72.32 ± 0.03%) and Glc (27.68 ± 0.03%). Anti-inflammatory assays results showed that as DNPs were 200 μg/mL, and the contents of NO, TNF-α, IL-1β, IL-6 and IL-10 in LPS-induced RAW 264.7 cells were about 13.39% and 13.39%, 43.88% and 43.51%, 17.80% and 15.37%, 13.84% and 20.66%, and 938.85% and 907.77% of those in control group, respectively. It was indicated that DNP1 and DNP2 inhibited the inflammatory response of RAW 264.7 cells induced by LPS via suppressing the level of NO and pro-inflammatory cytokines (TNF-α, IL-1β and IL-6) and promoting the secretion of anti-inflammatory cytokine (IL-10). Therefore, DNP1 and DNP2 have potential applications in the treatment of inflammatory injury.

Practical Qualitative Evaluation and Screening of Potential Biomarkers for Different Parts of Wolfiporia cocos Using Machine Learning and Network Pharmacology

Wolfiporia cocos is a widely used traditional Chinese medicine and dietary supplement. Artificial intelligence algorithms use different types of data based on the different strategies to complete multiple tasks such as search and discrimination, which has become a trend to be suitable for solving massive data analysis problems faced in network pharmacology research. In this study, we attempted to screen the potential biomarkers in different parts of W. cocos from the perspective of measurability and effectiveness based on fingerprint, machine learning, and network pharmacology. Based on the conclusions drawn from the results, we noted the following: (1) exploratory analysis results showed that differences between different parts were greater than those between different regions, and the partial least squares discriminant analysis and residual network models were excellent to identify Poria and Poriae cutis based on Fourier transform near-infrared spectroscopy spectra; (2) from the perspective of effectiveness, the results of network pharmacology showed that 11 components such as dehydropachymic acid and 16α-hydroxydehydrotrametenolic acid, and so on had high connectivity in the “component-target-pathway” network and were the main active components. (3) From a measurability perspective, through orthogonal partial least squares discriminant analysis and the variable importance projection > 1, it was confirmed that three components, namely, dehydrotrametenolic acid, poricoic acid A, and pachymic acid, were the main potential biomarkers based on high-performance liquid chromatography. (4) The content of the three components in Poria was significantly higher than that in Poriae cutis. (5) The integrated analysis showed that dehydrotrametenolic acid, poricoic acid A, and pachymic acid were the potential biomarkers for Poria and Poriae cutis. Overall, this approach provided a novel strategy to explore potential biomarkers with an explanation for the clinical application and reasonable development and utilization in Poria and Poriae cutis.

Soft Sensor Modeling Method Based on Improved KH-RBF Neural Network Bacteria Concentration in Marine Alkaline Protease Fermentation Process

Marine alkaline protease (MAP) fermentation is a complex multivariable, multi-coupled, and nonlinear process. Some unmeasured parameters will affect the quality of protease. Aiming at the problem that some parameters are difficult to be detected online, a soft sensing modeling method based on improved Krill Herd algorithm RBF neural network (LKH-RBFNN) is proposed in this paper. Based on the multi-parameter RBFNN model, the adaptive RBF neural network algorithm and control law are used to approximate the unknown parameters. The adaptive Levy flight strategy is used to improve the traditional Krill Herd algorithm, improve the global search ability of the algorithm, and avoid falling into local optimization. At the same time, the location update formula of Krill Herd algorithm is improved by using the calculation methods of similarity and agglomeration degree, and the parameters of adaptive RBFNN are optimized to improve its over correction and large amount of calculation. Finally, the soft sensing prediction model of bacterial concentration and relative active enzyme in map process based on LKH-RBFNN is established. The root mean square error and maximum absolute error of this model are 0.938 and 0.569, respectively, which are less than KH-RBFNN and PSO-RBFNN prediction models. It proves that the prediction error of LKH-RBFNN model is smaller and can meet the needs of online prediction of key parameters of map fermentation.