Applications of artificial intelligence in anaerobic co-digestion: Recent advances and prospects

Efficient dark-fermentation biohydrogen production by Shigella flexneri SPD1 from biowaste-derived sugars through green solvents approach

This study evaluated the dark-fermentative hydrogen (H2) production potential of isolated and identified Shigella flexneri SPD1 from various pure (glucose, fructose, sucrose, lactose, and galactose) and biowastes (coconut coir, cotton fiber, groundnut shells, rice-, and wheat-straws)-derived sugars. Among sugars, S. flexneri SPD1 exhibited high H2 production of up to 3.20 mol/mole of hexose using glucose (5.0 g/L). The pre-treatment of various biowastes using green solvents (choline chloride and lactic acid mixture) and enzymatic hydrolysis resulted in the generation of up to 36.0 g/L of sugars. The maximum H2 production is achieved up to 2.92 mol/mol of hexose using cotton-hydrolysate. Further, the upscaling of bioprocess up to 5 L of capacity resulted in a maximum yield of up to 3.06 mol/mol of hexose. These findings suggested that S. flexneri SPD1, a novel H2-producer, can be employed to develop a circular economy-based approach to produce clean energy.

Novel cofactor regeneration-based magnetic metal-organic framework for cascade enzymatic conversion of biomass-derived bioethanol to acetoin

Upgrading biomass-derived bioethanol to higher-order alcohols using conventional biotechnological approaches is challenging. Herein, a novel, magnetic metal-organic-framework-based cofactor regeneration system was developed using ethanol dehydrogenase (EtDH:D46G), NADH oxidase (NOX), formolase (FLS:L482S), and nicotinamide adenine dinucleotide (NAD+) for converting rice straw-derived bioethanol to acetoin. A magnetic zeolitic imidazolate framework-8@Fe3O4/NAD+ (ZIF-8@Fe3O4/NAD+) regeneration system for cell-free cascade reactions was introduced and used to encapsulate EtDH:D46G, NOX, and FLS:L482S (ENF). ZIF-8@Fe3O4/NAD+ENF created an efficient microenvironment for three-step enzyme cascades. Under the optimized conditions, the yield of acetoin from 100 mM bioethanol using ZIF-8@Fe3O4/NAD+ENF was 90.4 %. The regeneration system showed 97.1 % thermostability at 50 °C. The free enzymes retained only 16.3 % residual conversion, compared with 91.2 % for ZIF-8@Fe3O4/NAD+ENF after ten cycles. The magnetic metal-organic-framework-based cofactor regeneration system is suitable for enzymatic cascade biotransformations and can be extended to other cascade systems for potential biotechnological applications.

Machine learning modeling of thermally assisted biodrying process for municipal sludge

Preparation of activated carbons is an important way to utilize municipal sludge (MS) resources, while drying is a pretreatment method for making activated carbons from MS. In this study, machine learning techniques were used to develop moisture ratio (MR) and composting temperature (CT) prediction models for the thermally assisted biodrying process of MS. First, six machine learning (ML) models were used to construct the MR and CT prediction models, respectively. Then the hyperparameters of the ML models were optimized using the Bayesian optimization algorithm, and the prediction performances of these models after optimization were compared. Finally, the effect of each input feature on the model was also evaluated using SHapley Additive exPlanations (SHAP) analysis and Partial Dependence Plots (PDPs) analysis. The results showed that Gaussian process regression (GPR) was the best model for predicting MR and CT, with R2 of 0.9967 and 0.9958, respectively, and root mean square errors (RMSE) of 0.0059 and 0.354 ℃. In addition, graphical user interface software was developed to facilitate the use of the GPR model for predicting MR and CT by researchers and engineers. This study contributes to the rapid prediction, improvement, and optimization of MR and CT during thermally assisted biodrying of MS, and also provides valuable guidance for the dynamic regulation of the drying process.

Machine learning: an advancement in biochemical engineering

One of the most remarkable techniques recently introduced into the field of bioprocess engineering is machine learning. Bioprocess engineering has drawn much attention due to its vast application in different domains like biopharmaceuticals, fossil fuel alternatives, environmental remediation, and food and beverage industry, etc. However, due to their unpredictable mechanisms, they are very often challenging to optimize. Furthermore, biological systems are extremely complicated; hence, machine learning algorithms could potentially be utilized to improve and build new biotechnological processes. Gaining insight into the fundamental mathematical understanding of commonly used machine learning algorithms, including Support Vector Machine, Principal Component Analysis, Partial Least Squares and Reinforcement Learning, the present study aims to discuss various case studies related to the application of machine learning in bioprocess engineering. Recent advancements as well as challenges posed in this area along with their potential solutions are also presented.

Ensemble machine learning prediction of anaerobic co-digestion of manure and thermally pretreated harvest residues

This study aimed to clarify the statistical accuracy assessment approaches used in recent biogas prediction studies using state-of-the-art ensemble machine learning approach according to 10-fold cross-validation in 100 repetitions. Three thermally pretreated harvest residue types (maize stover, sunflower stalk and soybean straw) and manure were anaerobically co-digested, measuring biogas and methane yield alongside eight thermal preprocessing and biomass covariates. These were the inputs to an ensemble machine learning approach for biogas and methane yield prediction, employing three feature selection approaches. The Support Vector Machine prediction with the Recursive Feature Elimination resulted in the highest prediction accuracy, achieving the coefficient of determination of 0.820 and 0.823 for biogas and methane yield prediction, respectively. This study demonstrated an extreme dependency of prediction accuracy to input dataset properties, which could only be mitigated with ensemble machine learning and strongly suggested that the split-sample approach, often used in previous studies, should be avoided.

Machine learning-aided inverse design for biogas upgrading through biological CO2 conversion

The biogas upgrading process through bioconversion of CO2 to CH4 by hydrogenotrophic methanogens is an attractive strategy for energy decarbonation. Many studies have optimized operational parameters to improve key performance indicators such as CH4% and H2 utilization efficiency. However, inconsistent laboratory conditions make it challenging to compare results. Existing models for analyzing operating conditions can only assess the impact of individual conditions and lack the ability to simultaneously optimize multiple conditions. To address this, two XGBoost models were built with R2 of 0.779 and 0.903 with data collected from literatures and were embedded into multi-objective partitive swarm optimization algorithm to optimal operating conditions. Predictions were compared with experimental validations under optimized conditions, revealing an 8.50% and 2.95% relative error in CH4% and H2 conversion rate, respectively. This approach streamlines biogas upgrading processes, offering a data-driven solution to enhance efficiency and consistency in the pursuit of sustainable methane production.

Applying machine learning to anaerobic fermentation of waste sludge using two targeted modeling strategies

Anaerobic fermentation is an effective method to harvest volatile fatty acids (VFAs) from waste activated sludge (WAS). Accurately predicting and optimizing VFAs production is crucial for anaerobic fermentation engineering. In this study, we developed machine learning models using two innovative strategies to precisely predict the daily yield of VFAs in a laboratory anaerobic fermenter. Strategy-1 focuses on model interpretability to comprehend the influence of variables of interest on VFAs production, while Strategy-2 takes into account the cost of variable acquisition, making it more suitable for practical applications in prediction and optimization. The results showed that Support Vector Regression emerged as the most effective model in this study, with testing R2 values of 0.949 and 0.939 for the two strategies, respectively. We conducted feature importance analysis to identify the critical factors that influence VFAs production. Detailed explanations were provided using partial dependence plots and Shepley Additive Explanations analyses. To optimize VFAs production, we integrated the developed model with optimization algorithms, resulting in a maximum yield of 2997.282 mg/L. This value was 45.2 % higher than the average VFAs level in the operated fermenter. Our study offers valuable insights for predicting and optimizing VFAs production in sludge anaerobic fermentation, and it facilitates engineering practice in VFAs harvesting from WAS.

Machine learning-based prediction of methane production from lignocellulosic wastes

The biochemical methane potential test is a standard method to determine the biodegradability of lignocellulosic wastes (LWs) during anaerobic digestion (AD) with disadvantages of long experiment duration and high operating expense. This paper developed a machine learning model to predict the cumulative methane yield (CMY) using the data of 157 LWs regarding physicochemical characteristics, digestion condition and methane yield, with the coefficient of determination equal to 0.869. Model interpretability analyses underscored lignin content, organic loading, and nitrogen content as pivotal attributes for CMY prediction. For the feedstocks with a cellulose content exceeding about 50%, the CMY in the early AD stage would be relatively lower than those with low cellulose content, but prolonging digestion time could promote methane production. Besides, lignin content in feedstock surpassing 15% would significantly inhibit methane production. This work contributes to valuable guidance for feedstock selection and operation optimization for AD plants.

Real-time operation of municipal anaerobic digestion using an ensemble data mining framework

This study presents a novel approach for real-time operation of anaerobic digestion using an ensemble decision-making framework composed of weak learner data mining models. The framework utilises simple but practical features such as waste composition, added water and feeding volume to predict biogas yield and to generate an optimised weekly operation pattern to maximise biogas production and minimise operational costs. The effectiveness of this framework is validated through a real-world case study conducted in the UK. Comparative analysis with benchmark models demonstrates a significant improvement in prediction accuracy, increasing from the range of 50-80% with benchmark models to 91% with the proposed framework. The results also show the efficacy of the weekly operation pattern, which leads to a substantial 78% increase in biogas generation during the testing period. Moreover, the pattern contributes to a reduction of 71% in total days required for feeding and 30% in total days required for pre-feeding.

Zinc/carbon nanomaterials inhibit antibiotic resistance genes by affecting quorum sensing and microbial community in cattle manure production

This study used metagenomic sequencing to examine the effects of carbon-based zinc oxide nanoparticles (CZnONPs) and graphene-based zinc oxide nanoparticles (GZnONPs) on quorum sensing (QS), antibiotic resistance genes (ARGs) and microbial community changes during cattle manure production. The manure zinc content was significantly reduced in GZnONPs group. In the QS pathway, the autoinducer gene increases significantly in Control group, while the transporter and repressor genes experience a substantial increase in CZnONPs group. These results contributed to the significantly decreased the abundance of ARGs in GZnONPs group. The co-occurrence network analysis revealed a correlation between core ARGs and QS-related KEGG Orthology or ARGs' hosts, indicating that the selective pressure of zinc influences microbial QS, forming a unique ARG pattern in in vivo anaerobic fermentation. These findings suggest that implementing nutritional regulation in farming practices can serve as a preventive measure to mitigate the potential transmission of ARGs resulting from livestock waste.

Two-stage biohydrogen and methane production from sugarcane-based sugar and ethanol industrial wastes: A comprehensive review

The transition to renewable energy sources is crucial to ensure a sustainable future. Although the sugar and ethanol industries benefit from this transition, there are untapped opportunities to utilize the waste generated from the sugar and ethanol process chains through two-stage anaerobic digestion (TSAD). This review comprehensively discusses the utilization of various sugarcane-based industrial wastes by TSAD for sequential biohydrogen and methane production. Factors influencing TSAD process performance, including pH, temperature, hydraulic retention time, volatile fatty acids and alkalinity, nutrient imbalance, microbial population, and inhibitors, were discussed in detail. The potential of TSAD to reduce emissions of greenhouse gases is demonstrated. Recent findings, implications, and promising future research related to TSAD, including the integration of meta-omics approaches, gene manipulation and bioaugmentation, and application of artificial intelligence, are highlighted. The review can serve as important literature for the implementation, improvement, and advancements in TSAD research.

Prediction of composting maturity and identification of critical parameters for green waste compost using machine learning

Ensuring the maturity of green waste compost is crucial to composting processes and quality control of compost products. However, accurate prediction of green waste compost maturity remains a challenge, as there are limited computational methods available. This study aimed to address this issue by employing four machine learning models to predict two indicators of green waste compost maturity: seed germination index (GI) and T value. The four models were compared, and the Extra Trees algorithm exhibited the highest prediction accuracy with R2 values of 0.928 for GI and 0.957 for T value. To identify the interactions between critical parameters and compost maturity, The Pearson correlation matrix and Shapley Additive exPlanations (SHAP) analysis were conducted. Furthermore, the accuracy of the models was validated through compost validation experiments. These findings highlight the potential of applying machine learning algorithms to predict green waste compost maturity and optimise process regulation.

Microbial volatile organic compounds as novel indicators of anaerobic digestion instability: Potential and challenges

The wide application of anaerobic digestion (AD) technology is limited by process fluctuations. Thus, process monitoring based on screening state parameters as early warning indicators (EWI) is a top priority for AD facilities. However, predicting anaerobic digester stability based on such indicators is difficult, and their threshold values are uncertain, case-specific, and sometimes produce conflicting results. Thus, new EWI should be proposed to integrate microbial and metabolic information. These microbial volatile organic compounds (mVOCs) including alkanes, alkenes, alkynes, aromatic compounds are produced by microorganisms (bacteria, archaea and fungi), which might serve as a promising diagnostic tool for environmental monitoring. Moreover, mVOCs diffuse in both gas and liquid phases and are considered the language of intra kingdom microbial interactions. Herein, we highlight the potential of mVOCs as EWI for AD process instability, including discussions regarding characteristics and sources of mVOCs as well as sampling and determination methods. Furthermore, existing challenges must be addressed, before mVOCs profiling can be used as an early warning system for diagnosing AD process instability, such as mVOCs sampling, analysis and identification. Finally, we discuss the potential biotechnology applications of mVOCs and approaches to overcome the challenges regarding their application.

Machine Learning Approach for Determining and Optimizing Influential Factors of Biogas Production from Lignocellulosic Biomass

Machine learning (ML) was used to predict specific methane yields (SMY) with a dataset of 14 features from lignocellulosic biomass (LB) characteristics and operating conditions of completely mixed reactors under continuous feeding mode. The random forest (RF) model was best suited for predicting SMY with a coefficient of determination (R²) of 0.85 and root mean square error (RMSE) of 0.06. Biomass compositions greatly influenced SMYs from LB, and cellulose prevailed over lignin and biomass ratio as the most important feature. Impact of LB to manure ratio was assessed to optimize biogas production with the RF model. Under typical organic loading rates (OLR), optimum LB to manure ratio of 1:1 was identified. Experimental results confirmed influential factors revealed by the RF model and provided the highest SMY of 79.2% of the predicted value. Successful applications of ML for anaerobic digestion modelling and optimization specifically for LB were revealed in this work.

Artificial intelligence and machine learning for smart bioprocesses

In recent years, the digital transformation of bioprocesses, which focuses on interconnectivity, online monitoring, process automation, artificial intelligence (AI) and machine learning (ML), and real-time data acquisition, has gained considerable attention. AI can systematically analyze and forecast high-dimensional data obtained from the operating dynamics of bioprocess, allowing for precise control and synchronization of the process to improve performance and efficiency. Data-driven bioprocessing is a promising technology for tackling emerging challenges in bioprocesses, such as resource availability, parameter dimensionality, nonlinearity, risk mitigation, and complex metabolisms. This special issue entitled "Machine Learning for Smart Bioprocesses (MLSB-2022)" was conceptualized to incorporate some of the recent advances in applications of emerging tools such as ML and AI in bioprocesses. This VSI: MLSB-2022 contains 23 manuscripts, and summarizes the major findings that can serve as a valuable resource for researchers to learn major advances in applications of ML and AI in bioprocesses.

Tree-based machine learning model for visualizing complex relationships between biochar properties and anaerobic digestion

The ideal conditions for anaerobic digestion experiments with biochar addition are challenging to thoroughly study due to different experimental purposes. Therefore, three tree-based machine learning models were developed to depict the intricate connection between biochar properties and anaerobic digestion. For the methane yield and maximum methane production rate, the gradient boosting decision tree produced R² values of 0.84 and 0.69, respectively. According to feature analysis, digestion time and particle size had a substantial impact on the methane yield and production rate, respectively. When particle sizes were in the range of 0.3-0.5 mm and the specific surface area was approximately 290 m²/g, corresponding to a range of O content (>31%) and biochar addition (>20 g/L), the maximum promotion of methane yield and maximum methane production rate were attained. Therefore, this study presents new insights into the effects of biochar on anaerobic digestion through tree-based machine learning.