Dynamic modeling and optimization of cyanobacterial C-phycocyanin production process by artificial neural network

A kinetic-assisted growth curve prediction method for Chlamydomonas reinhardtii incorporating transfer learning

Traditional predictions of microalgal growth states rely on empirical or easily implementable kinetic models, leading to significant biases and elevated cost. This study proposes a kinetic-assisted machine learning method for predicting the growth curve of microalgal biomass under small sample conditions. Firstly, a microalgae growth kinetic model is constructed based on the logistic model. A two-stage kinetic fitting strategy is specified to account for the light-dark ratio. The Box-Behnken method is employed for experimental design. Then, using Two-stage TrAdaboost.R2 algorithm, the kinetic model is utilized as the source domain, and the experimental design data serves as the target domain for training machine learning models. The results indicate that the proposed method outperforms a single machine learning model in terms of prediction and has the potential to rapidly estimate microalgal growth trends under different conditions and accurately predict harvested biomass, potentially reducing the need for laborious, expensive, and time-consuming laboratory trials.

Enhancement of isoprene production in engineered Synechococcus elongatus UTEX 2973 by metabolic pathway inhibition and machine learning-based optimization strategy

An engineered Synechococcus elongatus UTEX 2973-IspS.IDI is used to enhance isoprene production through geranyl diphosphate synthase (CrtE) inhibition and process parameters (light intensity, NaHCO3 and growth temperature) optimization approach. A cumulative isoprene production of 1.21 mg/gDCW was achieved with productivity of 12.6 μg/gDCW/h in culture supplemented with 20 μg/mL alendronate. This inhibition strategy improvises the cumulative isoprene production 5.76-fold in presence of alendronate. The maximum cumulative production of isoprene is observed to be 5.22 and 6.20 mg/gDCW (54.4 and 64.6 μg/gDCW/h) at statistical and artificial neural network genetic algorithm (ANN-GA) optimized conditions, respectively. The overall increase of isoprene production is found to be 29.52-fold using an integrated approach of inhibition and ANN-GA optimization in comparison to unoptimized cultures without alendronate. This study reveals that alendronate use as a potential inhibitor and machine learning based optimization is a better approach in comparison to statistical optimization to enhance the isoprene production.

Machine learning in bioprocess development: from promise to practice

Fostered by novel analytical techniques, digitalization, and automation, modern bioprocess development provides large amounts of heterogeneous experimental data, containing valuable process information. In this context, data-driven methods like machine learning (ML) approaches have great potential to rationally explore large design spaces while exploiting experimental facilities most efficiently. Herein we demonstrate how ML methods have been applied so far in bioprocess development, especially in strain engineering and selection, bioprocess optimization, scale-up, monitoring, and control of bioprocesses. For each topic, we will highlight successful application cases, current challenges, and point out domains that can potentially benefit from technology transfer and further progress in the field of ML.

A transfer learning approach for predictive modeling of bioprocesses using small data

Predictive modeling of new biochemical systems with small data is a great challenge. To fill this gap, transfer learning, a subdomain of machine learning that serves to transfer knowledge from a generalized model to a more domain-specific model, provides a promising solution. While transfer learning has been used in natural language processing, image analysis, and chemical engineering fault detection, its application within biochemical engineering has not been systematically explored. In this study, we demonstrated the benefits of transfer learning when applied to predict dynamic behaviors of new biochemical processes. Two different case studies were presented to investigate the accuracy, reliability, and advantage of this innovative modeling approach. We thoroughly discussed the different transfer learning strategies and the effects of topology on transfer learning, comparing the performance of the transfer learning models against benchmark kinetic and data-driven models. Furthermore, strong connections between the underlying process mechanism and the transfer learning model's optimal structure were highlighted, suggesting the interpretability of transfer learning to enable more accurate prediction than a naive data-driven modeling approach. Therefore, this study shows a novel approach to effectively combining data from different resources for bioprocess simulation.

Application of a GA-Optimized NNARX controller to nonlinear chemical and biochemical processes

Chemical and biochemical processes generally suffer from extreme nonlinearities with respect to internal states, manipulated variables, and also disturbances. These processes have always received special technical and scientific attention due to their importance as the means of large-scale production of chemicals, pharmaceuticals, and biologically active agents. In this work, a general-purpose genetic algorithm (GA)-optimized neural network (NNARX) controller is introduced, which offers a very simple but efficient design. First, the proof of the controller stability is presented, which indicates that the controller is bounded-input bounded-output (BIBO) stable under simple conditions. Then the controller was tested for setpoint tracking, handling modeling error, and disturbance rejection on two nonlinear processes that is, a continuous fermentation and a continuous pH neutralization process. Compared to a conventional proportional-integral (PI) controller, the results indicated better performance of the controller for setpoint tracking and acceptable action for disturbance rejection. Hence, the GA-optimized NNARX controller can be implemented for a variety of nonlinear multi-input multi-output (MIMO) systems with minimal a-priori information of the process and the controller structure.

Bioprocess Control: Current Progress and Future Perspectives

Typical bioprocess comprises of different unit operations wherein a near optimal environment is required for cells to grow, divide, and synthesize the desired product. However, bioprocess control caters to unique challenges that arise due to non-linearity, variability, and complexity of biotech processes. This article presents a review of modern control strategies employed in bioprocessing. Conventional control strategies (open loop, closed loop) along with modern control schemes such as fuzzy logic, model predictive control, adaptive control and neural network-based control are illustrated, and their effectiveness is highlighted. Furthermore, it is elucidated that bioprocess control is more than just automation, and includes aspects such as system architecture, software applications, hardware, and interfaces, all of which are optimized and compiled as per demand. This needs to be accomplished while keeping process requirement, production cost, market value of product, regulatory constraints, and data acquisition requirements in our purview. This article aims to offer an overview of the current best practices in bioprocess control, monitoring, and automation.

Computational Analysis of Dynamic Light Exposure of Unicellular Algal Cells in a Flat-Panel Photobioreactor to Support Light-Induced CO2 Bioprocess Development

Cyanobacterial cell factories trace a vibrant pathway to climate change neutrality and sustainable development owing to their ability to turn carbon dioxide-rich waste into a broad portfolio of renewable compounds, which are deemed valuable in green chemistry cross-sectorial applications. Cell factory design requires to define the optimal operational and cultivation conditions. The paramount parameter in biomass cultivation in photobioreactors is the light intensity since it impacts cellular physiology and productivity. Our modeling framework provides a basis for the predictive control of light-limited, light-saturated, and light-inhibited growth of the Synechocystis sp. PCC 6803 model organism in a flat-panel photobioreactor. The model here presented couples computational fluid dynamics, light transmission, kinetic modeling, and the reconstruction of single cell trajectories in differently irradiated areas of the photobioreactor to relate key physiological parameters to the multi-faceted processes occurring in the cultivation environment. Furthermore, our analysis highlights the need for properly constraining the model with decisive qualitative and quantitative data related to light calibration and light measurements both at the inlet and outlet of the photobioreactor in order to boost the accuracy and extrapolation capabilities of the model.

Engineering salt tolerance of photosynthetic cyanobacteria for seawater utilization

Photosynthetic cyanobacteria are capable of utilizing sunlight and CO₂ as sole energy and carbon sources, respectively. With genetically modified cyanobacteria being used as a promising chassis to produce various biofuels and chemicals in recent years, future large-scale cultivation of cyanobacteria would have to be performed in seawater, since freshwater supplies of the earth are very limiting. However, high concentration of salt is known to inhibit the growth of cyanobacteria. This review aims at comparing the mechanisms that different cyanobacteria respond to salt stress, and then summarizing various strategies of developing salt-tolerant cyanobacteria for seawater cultivation, including the utilization of halotolerant cyanobacteria and the engineering of salt-tolerant freshwater cyanobacteria. In addition, the challenges and potential strategies related to further improving salt tolerance in cyanobacteria are also discussed.

Harnessing the potential of artificial neural networks for predicting protein glycosylation

Kinetic models offer incomparable insight on cellular mechanisms controlling protein glycosylation. However, their ability to reproduce site-specific glycoform distributions depends on accurate estimation of a large number of protein-specific kinetic parameters and prior knowledge of enzyme and transport protein levels in the Golgi membrane. Herein we propose an artificial neural network (ANN) for protein glycosylation and apply this to four recombinant glycoproteins produced in Chinese hamster ovary (CHO) cells, two monoclonal antibodies and two fusion proteins. We demonstrate that the ANN model accurately predicts site-specific glycoform distributions of up to eighteen glycan species with an average absolute error of 1.1%, correctly reproducing the effect of metabolic perturbations as part of a hybrid, kinetic/ANN, glycosylation model (HyGlycoM), as well as the impact of manganese supplementation and glycosyltransferase knock out experiments as a stand-alone machine learning algorithm. These results showcase the potential of machine learning and hybrid approaches for rapidly developing performance-driven models of protein glycosylation.

Phototrophic cultivation of NaCl-tolerant mutant of Spirulina platensis for enhanced C-phycocyanin production under optimized culture conditions and its dynamic modeling

Commercial cultivation of Spirulina sp. is highly popular due to the presence of high amount of C-phycocyanin (C-PC) and other valuable chemicals like carotenoids and γ-linolenic acid. In this study, the pH and the concentrations of nitrogen and carbon source were manipulated to achieve improved cell growth and C-PC production in NaCl-tolerant mutant of Spirulina platensis. In this study, highest C-PC (147 mg · L^-1 ) and biomass (2.83 g · L^-1 ) production was achieved when a NaCl-tolerant mutant of S. platensis was cultivated in a nitrate and bicarbonate sufficient medium (40 and 60 mM, respectively) at pH 9.0 under phototrophic conditions. Kinetic study of wildtype S. platensis and its NaCl-tolerant mutant was also done to determine optimum nitrate concentrations for maximum growth and C-PC production. Kinetic parameter of inhibition (Haldane model) was fitted to the relationship between specific growth rate and substrate concentration obtained from the growth curves. Results showed that the maximum specific growth rate (μ_max ) for NaCl-tolerant mutant increased by 17.94% as compared to its wildtype counterpart, with a slight increase in half-saturation constant (K_s ), indicating that this strain could grow well at high concentration of NaNO₃ . C-PC production rate (C_max ) in mutant cells increased by 12.2% at almost half the value of K_s as compared to its wildtype counterpart. Moreover, the inhibition constant (K_i ) value was 207.85% higher in NaCl-tolerant mutant as compared to its wildtype strain, suggesting its ability to produce C-PC even at high concentrations of NaNO₃ .

An efficient model construction strategy to simulate microalgal lutein photo-production dynamic process

Lutein is a high-value bioproduct synthesized by microalga Desmodesmus sp. It has great potential for the food, cosmetics, and pharmaceutical industries. However, in order to enhance its productivity and to fulfil its ever-increasing global market demand, it is vital to construct accurate models capable of simulating the entire behavior of the complicated dynamics of the underlying biosystem. To this aim, in this study two highly robust artificial neural networks (ANNs) are designed for the first time. Contrary to conventional ANNs, these networks model the rate of change of the dynamic system, which makes them highly relevant in practice. Different strategies are incorporated into the current research to guarantee the accuracy of the constructed models, which include determining the optimal network structure through a hyper-parameter selection framework, generating significant amounts of artificial data sets by embedding random noise of appropriate size, and rescaling model inputs through standardization. Based on experimental verification, the high accuracy and great predictive power of the current models for long-term dynamic bioprocess simulation in both real-time and offline frameworks are thoroughly demonstrated. This research, therefore, paves the way to significantly facilitate the future investigation of lutein bioproduction process control and optimization. In addition, the model construction strategy developed in this research has great potential to be directly applied to other bioprocesses. Biotechnol. Bioeng. 2017;114: 2518-2527. © 2017 Wiley Periodicals, Inc.