The role of statistical methods in optimizing and enhancing fungal chitosan commercial production

Abstract

This review portrays the role of fermentation technology in the production of fungal chitosan (FCH) under optimized conditions using statistical  methods. It is noteworthy to mention that FCH is superior than crustacean chitosan (CCH) due to its low molecular weight (LMW) (≈ 20–30 kDa), polymer homogeneity, high degree of deacetylation (DDA) (74–92%), along with thermal stability, solubility at wide physiological pH and greener extraction process. Employment of suitable submerged fermentation conditions improves the quality (high DDA and LMW) of FCH for varied applications. Literature survey depicted the crucial role of recent advancements of statistical tools and software in FCH fermentation technology. A close look at the literature over the past three decades showed ≈ 64% of FCH production from Absidia coerulea, Rhizopus oryzae, R. japonicus, Aspergillus niger, A. terreus, A. flavus, Cunninghamella elegans (≈ 16% each) followed by Mucor rouxii (≈ 11%). Other fungi namely, Benjaminiella poitrasii, Penicillium chrysogenum and Trametes versicolor, Gongronella butleri and Ganoderma lucidum (≈ 5% each) have been reported. The Design of Experiments (DOE), like response surface methodology (RSM) including Plackett–Burman Design (PBD), Central composite design (CCD), Box Behnken design (BBD) and Taguchi have improved biomass and FCH yield meaningfully. Among different approaches, One-factor-at-a-time (OFAT) approach was the foremost choice (≈ 29%) followed by CCD (≈ 12%) and OFAT combined with CCD (≈11%) were employed by researchers to optimize FCH production from potent strains. Around 6% of the reports suggest that BBD, Taguchi, FC-BBD, CCD, 2 > 2 factorials have been employed at an individual level to achieve a high yield of FCH. Those methods can be employed either individually or in combination. This article comprehensively presents the basic information, performances of the statistical   methods/tools of DOE and software employed for successful scaling-up of FCH while highlighting their merits, limitations, and challenges.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13205-025-04236-2.

Introduction

Several essential economic products such as vaccines, drugs, textiles, food, feed, chemicals and energy are produced on large-scale using fermentation technology (Ahmed et al. 2023; Sun et al. 2022; BT et al. 2019; Abd-Elhalem et al. 2015). There has been significant discussion in recent years regarding enhancing yields of desired products through optimizing fermentation conditions. Established literature has shown that the most common theory of optimization dates to 100 BC which was employed to measure distances between two objects or any two specified points (Sinha 2019). Although fermentation technologies have been utilized for decades, the search for sustainable production that meets market demands while being cost-effective has always been challenging (Abd-Elhalim et al. 2023; Singh et al. 2017). Optimizing key factors of the manufacturing process including appropriate media components, their suitable concentrations, biomasses and metabolite productivity are crucial in achieving maximum products’ yields (Rajitha and Naik 2021). Various studies have promoted the composition of fermentation media as a major contributor towards enhancing the yield of product/metabolite and volumetric productivity as crucial factors leading to feasibility in commercial fermentation processes (Abd-Elhalim et al. 2023; Drożdżyńska et al. 2023). In addition, the cost of both the fermentation medium and the downstream processing (40 to 60%) contributes to the process economics (Kennedy and Krouse 1999; Gupta et al. 2023). The goals of the optimization processes always aim to improve and maximize the yield of fermentative products along with the efficiency of each process within a sustainable framework (Singh et al. 2017).

The optimization process, however, may face several challenges and hence lead to significant concerns among researchers. A single design cannot fully satisfy all the desirable parameters demanded by the fermentation processes. Kennedy and Krouse (1999) aptly described this by stating that “One magic bullet technique for all situations does not exist and that every technique will have its own set of advantages and disadvantages”. Researchers assume that when a media is referred to as optimized, it is simply the ‘best-till-now’ and not the perfect composition (Kennedy and Krouse 1999). This solicits the question that “Is optimization to be performed from scratch every time when a new process is undertaken?” Nonetheless, this fact is certainly not completely TRUE in most cases, since prior research and literature offer eloquent insights and an overall advantage that can be adopted effectively in these techniques. Until the 1970s, classical and conventional approaches such as one-factor-at-a-time (OFAT), were generally used for identifying the media components suitable for the fermentation process. Moreover, the use of OFAT is an expensive, laborious and time-consuming approach. The evolution of statistical experimental strategies such as the Design of Experiment (DOE) approach has lately become more competent and economically feasible as a sophisticated, cost-effective, and influential tool to evaluate productivity and quality enhancement instead of relying on trial-and-error approaches (Kumar et al. 2016).

The DOE-based optimization methodology has become indispensable in numerous industrial sectors as an integral element of decision-making processes either for new product development/manufacturing processes or improvement of existing processes (Bonaccorso et al. 2021). DOE must first identify parameters that affect a bioprocess. The screening is required generally to select the parameters that significantly influence the said process. The use of statistical models or designs viz., Response Surface Methodology (RSM), Artificial Intelligence (AI)-based advanced systems such as Genetic Algorithms (GA), and Artificial Neural Networks (ANN) are helpful in determining the components concentrations of fermentation media along with other process parameters (Lau et al. 2023). Prior to using a statistical tool, it is critical to select valid mathematical models by comparing or testing them objectively in small-scale experiments verifying their appropriateness. A study by Ding et al. (2021) illustrated subjective testing in order to rank their competing models and determine their advantages and limitations. The approach used was adapted to screen and improve image quality assessment models. The approach, however, can be chosen for screening statistical designs for optimizing several fermentation processes.

Sopacua et al. (2023) reviewed the application of RSM in the fermentation of various food products. Appropriate use of RSM was helpful in employing statistical experimental designs to discover the optimal value of any response. These optimization techniques have been implemented for almost every biological process including enzymes, metabolites, and bioactive compounds from plants and microorganisms. Qadir et al. (2022) demonstrated the effect of enzyme complexes on the recovery of phenolics from the fruit extract of Capparis spinosa by employing RSM and ANN approaches. The treatment conditions (enzyme concentration, pH, temperature, and time) for recovery of phenolics from C. spinosa fruits were optimized. This article particularly deals with chitosan in general, crustacean and fungal-origin polymer. Abbreviations for the general term—chitosan (CH), crustacean chitosan (CCH) and fungal chitosan (FCH) will be used henceforth throughout the article. CH, a multifarious deacetylated derivative of chitin (CT) extracted from crustacean, microbial sources is of immense interest for diverse industrial applications (Karamchandani et al. 2024a; 2024b; Kou et al. 2021; Kean and Thanou 2011). CT is the second most abundant biopolymer present on earth and its N-deacetylated form is known as CH. The CH has one primary amine (- NH₂) and two free hydroxyl (- OH) groups. At the C-2 position, (- OH) in CH has been replaced by acetamido or amino groups (Karamchandani et al. 2024b, 2024c). CH is a polysaccharide that consists of β‐1,4‐GlcNAc and β‐1, 4‐GlcN linkages and obtained from CT by partial or complete alkali treatment (Karamchandani et al. 2022a). The most common sources for CCH production are marine animals such as shrimps and crabs. Whereas, FCH is extracted from diverse fungal communities. The production of FCH is superlative as compared to CCH since it has a low molecular weight (low ≈ 20–30 kDa), homogeneity in polymer length, high degree of deacetylation (DDA:74–92%) along with reasonable thermal stability and a simpler extraction process for the manufacturers. Production of CCH from marine shell sources is not an environmentally friendly process due to the requirement of extensive chemical treatments (demineralization, deacetylation and deproteinization processes). The growing concern about the use of ecologically friendly biomaterials has encouraged researchers to explore ‘Zygomycetes filamentous fungi’ as an alternative source of CT. Thus, extraction of FCH from those fungi offers a suitable way to produce an eco-friendly approach. Further advantages like no dependency on season, site and low-cost requirement for the management of waste and demineralization process are attractive economic features to encourage manufacturers worldwide (Karamchandani et al. 2024b; Ghormade et al. 2017). The commercial scale production (≈ 30–50%) of FCH has been solely reported from the Zygomycetes group of Cunninghamellaceae and Mucoraceae families. However, it is important to identify factors or parameters that could facilitate improvement in FCH yield.

The quantity and characteristics of FCH are undoubtedly influenced by factors like the producing microbial strain, fermentation medium, growth conditions and physicochemical parameters. Therefore, it is mandatory to optimize these parameters to scale up FCH production from leading strains (Karamchandani et al. 2024c). This review presents systematic information on several approaches namely, One-Factor-At-a-Time (OFAT), Plackett–Burman Design (PBD), Taguchi method, Factorial design and RSM, which are typically employed to achieve higher yields of FCH. Various optimization studies, advancement of mathematical tools and software like Design-Expert which provide new dimensions for optimization of FCH production are presented in this review. The efficient use of several optimization techniques, either singly or in combination has been suggested to produce FCH in high quantity from the desired fungi. This review provides recommendations for appropriate analysis of data and experimental models (employing Design-Expert by Stat-ease) for optimum FCH production. Finally, it also covers the industrial point of view for FCH optimization and aims to confirm the most reliable tools to statistically design experiments and execute optimization with greater reliability.

The mechanism of chitosan synthesis in fungal cell wall

Fungal cell walls are composed of polysaccharides including glucan cross-linked CT fibres that strengthen it. Fungi are classified into four classes namely Basidiomycetes, Ascomycetes, Zygomycetes, and Deuteromycetes among which Zygomycetes bear a distinctive quantity of CH and CT in their cell wall (Gholizadeh and Aghdam 2010). Zygomycetes fungi have an extensive elongated mycelium network with varied structures of asexual and sexual spores. The vegetative mycelial structures of Zygomycetes do not comprise septa except in the region where it is needed for parting spores (Kirk et al. 2008). Zygomycetes fungi grow exceptionally fast as compared to others possibly due to the absence of septa providing an opportunity for quicker movement or intra-hyphal cytoplasmic transportation (Pochanavanich and Suntornsuk 2002). Fungal mycelial cell walls are composed of major structural and interstitial components namely polysaccharides (CT and glucan) and glycoproteins (mannoproteins, galactoproteins, xylomannoproteins and glucuronoproteins).

Chitosan, an amino-polysaccharide has significant biocompatibility, biodegradability, non-toxicity and exceptional antimicrobial properties (Abo Elsoud et al. 2023; Muzzarelli 2010). CT–a crystalline microfibril is a precursor for forming a valuable and reactive form polymer i. e. CH (Kou et al. 2021; Kean and Thanou 2011). Deacetylation of the unit N-acetyl-D-glucosamine results in β−1, 4-D-glucosamine with repeating units in CH molecules. This allows better solubility of CH in different solvents compared to CT. Thus, the solubility of CH enhances its wide usage in agriculture, biomedical, cosmetics, and paper industrial sectors (Kou et al. 2021; Younes and Rinaudo 2015; Kean and Thanou 2011). The presence of abundant quantities of CH conceivably works as a defence system to provide protection from hydrolytic attacks by chitinases enzymes (Synowiecki and Al-Khateeb 2003). CT production in fungi involves polymerization of glucosamine (GlcNAc) via an enzymatic pathway. In the fungal cell wall, the polymer N-acetylglucosamine is synthesised from the primary source (either glucose or trehalose) where 1. GlcNAc (amino sugar) is synthesized, followed by 2. formation of the UDP-GlcNAc (activation step). First and second-step reactions occur in the cytoplasm of the fungus. Finally, 3. The polymerization of CT is mediated by UDP-GlcNAc (which acts as the activated donor of sugar) and take place in the plasma membrane. In the final step, CT in the presence of chitin deacetylase enzyme (CDA) executes the deacetylation of GlcNAc leading to the formation of FCH (Huq et al. 2022; Crognale et al. 2022) (Fig. 1). Thus, the CDA enzyme activity is crucial for the synthesis of FCH in the fungal cell wall.

Fig. 1.

Biochemical mechanism of synthesis of chitosan in the cell wall of Zygomycetes fungi. Index on the left-hand side indicates different components with different colour codes involved in the synthesis of the fungal cell wall. The enzymatic synthesis of the fungal chitosan (FCH) begins in the cytoplasm with the production of N-acetylglucosamine (GlcNAc) from glucose or trehalose, followed by its activation to form uridine diphosphate-N-acetylglucosamine (UDP-GlcNAc) by an enzyme chitin synthase (CS slipper shaped dark orange colour). In the plasma membrane (phospholipid bilayer), UDP-GlcNAc  (small rod-shaped light brown colour) - a activated sugar donor is used for the polymerization of chitin (CT dumbbell-shaped brown colour). Finally, the enzyme chitin deacetylase (CDA half purple colour circle) catalyzes the conversion of CT into FCH (CH dumbbell-shaped white colour with brown border) in the cell wall. (Figure constructed using Bio render software, Professional Science Figure creator, BioRender, Toronto, ON, Canada)

The molecules of CH in the fungal cell wall co-exist in two different forms, either as free amino-glucosides or covalently bonded to β-glucan (Raafat and Sahl 2009). The characteristic properties of CH as a component of fungal cell wall include higher integrity, protection from high temperatures and exposure to various cell inhibitors (Banks et al. 2005; Adams 2004). In nature, both CT and CH biomaterials are not found in the cell wall of all fungal species. In contrast, the ‘Zygomycetes fungi’ shows the presence of higher percentages of CT and CH in their cell wall and therefore are a rich resource to extract these valuable biomaterials in economical ways (Azeez et al. 2023). The quantity and properties of FCH are certainly affected by several factors (such as fungal strain, media components, physicochemical growth parameters) and, therefore, optimized conditions are essential to achieve high yields through improved scale-up procedures.

Several Zygomycetes fungi viz. Cunninghamella blakesleeana, C. echinulata, Absidia coerulea, Benjaminiella poitrasii, Gongronella butleri, Mucor javanicus and Rhizopus oryzae have been reported for FCH production and its applications in diverse fields (Ahmed et al. 2023; Nwe et al. 2010). Additionally, to-date, many other novel fungal strains have been reported for FCH extraction. For example, Auricularia sp. has been explored for FCH production along with its antibacterial activity against both Gram-positive and Gram-negative bacteria. The antibacterial activity of FCH was remarkable compared to the commercial CCH (Chang et al. 2019). The researchers further reported that the DDA of extracted FCH was 86.81%. The study by El Rabey et al. (2019) showed the production of FCH from a novel fungal strain Amylomyces rouxii where biopolymer had high DDA (88.7%). Such FCH has promising applications in health care disciplines. Researchers suggested enhanced antimycotic activity of the finished cotton-based textile with Fluconazole/chitosan nanoparticles nanoconjugates against two strains of C. albicans. Media composition optimized along with cultivation conditions (physiological : pH, temperature, agitation, aeration) are required in enhancing fungal biomass production along with FCH yield. Few protocols are used to extract FCH from the potent fungi. However, a two-step extraction procedure comprising acid and base treatment is a simple one and the most popular in the research community and is discussed in the following sections.

A two-step extraction procedure (acid and base treatment) for extraction of FCH

Several researchers have employed the two-step extraction protocol to capture FCH from the cell of potent fungi (Karamchandani et al. 2022b; Pochanavanich and Suntornsuk 2002). In this method, the protocol starts with the preparation of a fungal spore suspension (1.0 × 10⁸) in a phosphate buffer saline (PBS) and then further inoculated in a selected fermentation broth where it is allowed to grow. After completing the fermentation process, the fungal biomass is separated by centrifugation. Subsequently, the fungal biomass is harvested by filtration technique using Whatman filter paper (Grade No.1) followed by alkaline treatment (1 N NaOH 1:30 w/v) to separate proteins and glucans from the supernatant. The solution is then autoclaved for 15 min at 121 °C before further treatment. The alkali-insoluble material is pelleted by centrifugation at 10,000 rpm for 15 min followed by the addition of distilled water to maintain the neutral pH. The obtained substance was then dried in a hot air oven at 60 °C until constant weight was achieved. The insoluble mass obtained is treated with acetic acid (1% v/v) and the mixture is then maintained in a hot water bath at 95 °C for 5 h. This imperative step disassembles the FCH components from the fungal cell wall which is centrifuged at 10,000 rpm for 20 min. The supernatant obtained is then mixed with 2 N NaOH to make the pH of the solution up to 10. The obtained material is then subjected to another round of centrifugation (10,000 rpm for 20 min) after which the pellet is collected. This extracted FCH is washed multiple times with distilled water followed by treatment with 95% of ethanol. Finally, the precipitate is then washed with acetone and dried in a hot air oven at 60 °C until the homogenized weight is achieved. The extracted FCH could be stored at 4 °C to be used for future applications. Recently, a report on the extraction of FCH from C. echinulata NCIM 691 has been published by our group (Karamchandani et al. 2024a; 2024b; 2024c; 2022b).

The role of various media components and culture conditions in enhancing fungal biomass and its chitosan production

Media composition plays a crucial role in microbial fermentation technology. As previously discussed, the fungal cell wall is composed of CT, CH, glycoproteins, α−1,3-glucan, β−1,3-glucan, β−1,6-glucan chains and mannans (Free 2013). Consequently, fungi require exact carbon (C) and nitrogen (N) sources in appropriate concentrations to synthesize proteins, carbohydrates, amino acids, nucleotides, cellular components and metabolites that are essential for their growth. Several physiological, metabolic activities including synthesis of various ‘N’ containing polysaccharides like CT and CH (Zhang and Elser 2017) are affected by the conditions provided for the growth of microbes. The elemental composition of fungal biomass shows around 38 to 57% of ‘C’, 0.23 to 15% of ‘N’, and 0.040 to 5.5% phosphorus (P). On an average, the C:N:P ratio in the fungal cell wall is around 250:16:1. Gao et al. (2007) observed great variation in the requirements for C:N ratio by fungal species which indicated that the requirements of ‘C’,‘N’ sources and C:N ratio for fungal growth and sporulation could be strain-dependent. Thus, while designing an optimal medium, concentrations of ‘C’, ‘N’ etc. are also critical factors. Camenzind et al. (2021) demonstrated that varying fungal C:N:P ratios are virtuous linear predictors of fungal growth and activity when the growth media has ‘N’ and ‘P’ in limited quantities. Di Lonardo et al. (2020) investigated the effect of different concentrations of ‘N’ source on the growth of Trichoderma harzianum and M. hiemalis. The authors concluded that in spite of the lower C:N ratio in media composition, both fungi could grow luxuriously. The optimum and appropriate ratios of C:N ensures that ‘C’ source is not easily exhausted within a short period of time before ‘N’ or other nutrients reduce the yield of fungal biomass and CH. It is essential to optimize the concentration of ‘C’ with respect to ‘N’ without hindering the enzymatic activity by the media components especially during the optimization of fungal growth process. Hence while designing a media, one should aim in attaining high biomass and FCH yield. The use of standardized commercially available dehydrated media like potato dextrose broth / potato dextrose agar (PDB/PDA) (containing potato infusion and dextrose), yeast peptone glucose / yeast peptone fructose (YPG/YPF) have been proven to be effective for fungal growth and production of FCH (Chatterjee et al. 2005; Fang and Zhong 2002). Other than dehydrated media, many more cheap and renewable substrates like molasses, apple waste, potato peels, and mushroom waste have supported FCH production (Habibi et al. 2021; Wu et al. 2004). Additionally, physicochemical parameters such as temperature, pH, aeration, and agitation speed are decisive and are discussed in the following sections.

Significance of carbon and nitrogen in fermentation technology

Carbon is the key factor in the fermentation process of any microbial metabolite. The cost of raw materials is also equally important along with the rate of assimilation of a particular type of ‘C’ source by a given organism. Different ‘C’ sources such as monosaccharides (glucose, fructose, mannitol, galactose, etc.) or disaccharides (sucrose, maltose, sucrose, etc.) influence the quantity of microbial biomass. Similarly, organic nitrogen sources like tryptone, peptone, beef extract, yeast extract and inorganic sources such as ammonium salts and nitrates are also important factors responsible for determining biomass content. Fungi do exhibit interesting preference patterns towards utilization of varied ‘C’ and ‘N’ sources during the production of FCH and their biomass. Wang et al. (2008) recorded the autolysis in A. coerulea after three days when the culture was initially grown on a medium supplemented with glucose or fructose. However, when the fermentation medium was replaced with sucrose, the mycelial growth slowly progressed until day seven of incubation and resulted in a better yield of low molecular weight FCH (LMWFCH). Thus, in the presence of sucrose A. coerulea generated high yield of FCH followed by molasses, maltose, fructose and glucose.

Habibi et al. (2021) analysed the effect of ‘C’ source on the growth of A. terreus in a medium provided with apple juice at different concentrations (20, 40, 60, and 80 w/v %) while ‘N’ source (ammonium nitrate) was kept constant (3 g/l). They reported that when the apple extract concentration was increased from 20 to 80 (w/v %), fungal cell growth was enhanced indicating that the carbohydrates present in the apple extract supported A. terreus growth. Biomass of Pleurotus ostreatus produced optimal amounts of 175.92 and 3.7 mg/g of CT and FCH, respectively when grown in a medium containing a mixture of agricultural and agro-industrial residues like olive pomace, wheat straw and coffee grounds (Benamar et al. 2022). Another example is the cultivation of Penicillium expansum, A. niger, R. oryzae and Fusarium moniliforme which have been reported for FCH production on substrates like corn straw (Omogbai and Ikenebomeh 2013). Recently, corn steep liquor (CSL) and cassava wastewater (‘C’, ‘N’ sources) were shown to affect the morphology, biomass yields and FCH production by M. subtilissimus UCP 1262 and Lichtheimia hyalospora UCP 1266 (de Souza et al. 2020).

Impact of changes in process parameters

Several cultivation parameters such as temperature, initial pH, inoculum age, size, and aeration conditions affect fungal CT and CH production. Under controlled laboratory conditions, several fungi generally grow at temperatures, pH and agitation ranging between 25–30 °C, 3–6 and 100–250 rpm, respectively. An improved cell growth and CT productivity has been achieved by fungus Ganoderma lucidum through the optimization of pH conditions and inoculum size (Mokhtari-Hosseini et al. 2018). Authors have reported optimum conditions of 5.7 pH, 7.45% inoculum size to produce 5.91 g/l of cell dry weight and 1.08 g/l of CT. Similar work carried out by Kim et al. (2001) investigated the effects of varying pH on microbial growth and FCH production by A. coerulea (ATCC 14076). Using similar media components (20 g/l: glucose, 15 g/l: yeast extract, 5 g/l: ammonium sulphate), altered pH conditions (reduced from 6.3 to 4.8) affected the cell growth and thereby altered the biomass productivity. When the same strain was cultivated at a lower pH (4.5), it resulted in a maximum yield of (2.3 g/l) FCH after 36 h. The same experiment was continued further to evaluate the effect of agitation on the yield of FCH. The optimum agitation speed for cell growth was 250 rpm, whereas lowering the agitation speed to 150 rpm reduced the yield of FCH which could be probably due to the autolysis of the cells affecting the rate of oxygen and mass transfer.

The above reports confirmed the importance of the various process variables on FCH production. To augment the fungal biomass production and enhance the yield of FCH; designing media and optimization of its components is indispensable. A variety of techniques including conventional and advanced approaches employed by the researchers in optimizing the FCH yield are described as follows.

Employing suitable statistical methods / models for optimization

Conventional approach: one-factor-at-a-time (OFAT)

This is the most conventional method used so far by many researchers to establish the role of key factors involved in optimizing the production of metabolites from elite microbial cultures. However, selecting One-Factor-At-A-Time (OFAT) is advantageous at the initial stages and is based on three key principles viz., 1. Removal, 2. Supplementation and 3. Replacement (Singh et al. 2017). In these three steps, the removal of less important factor/s could be carried out followed by the addition of new or required factors. While working only on one factor, replacement of certain parameters is possible for considering the economic feasibility of the intended metabolite or the product. Basically, OFAT is a screening tool that unveils the media components essential for preliminary experimental set-up. In the OFAT technique, the results are dependent on the collection of data after each trial run and also comparing it with previous outputs. Hence, this method is not suitable for the analysis of a large number of variables. Since there is no statistical base for the OFAT method, it is a trial-and-error technique (Wahid and Nadir 2013) and to overcome the flaws of the method, there is a greater need to emphasize statistical approaches for designing optimal systems (Frey and Jugulum 2006).

Statistical approaches

The process of designing production media for the microbial fermentation process can be expensive and laborious. Applying Design of Experiment (DOE) methods involving the usage of statistical approaches reduces not only the number of experiments but also the time and resources required to achieve a high yield of microbial products (Keskin Gündoğdu et al. 2016). The introduction of statistical tools in microbial technology aids in determining the vital variables that could be controlled while eliminating unimportant variables (Vanaja and Shobha Rani 2007). These statistical methods aid in extrapolating data and identify the best possible ranges and combinations of the test variables through planning, conducting, and analysing data obtained from experiments (Rodriguez-Granrose et al. 2021). The justification can be supported by representing graphs to depict the valuable relationship between the variables and obtaining optimum yield of the product. Unlike OFAT,  these methods doesn’t have limitations like time requirement, inaccuracy or missing any important information (Frey and Jugulum 2006). This is applicable in designing not only new products but also in formulation, process improvement, developing manufacturing process and process commercialization in robust ways.

Hence, DOE facilitates a better understanding of the process in identifying the critical parameters involved in the high production of intended microbial products. Therefore, DOE is a well-appreciated technique due to the following reasons (i) involving few numbers of experiments, (ii) low cost, (iii) less quantity of raw materials and (iv) less time to acquire high quality information. Organizing the experiments through DOE approaches is always helpful to achieve the proposed objectives through the planning of accurate protocols. The DOE approach may comprise a ‘Full Factorial Design’ where all factors (medium components, pH, temperature, agitation, etc.) can be considered in combination to properly execute all intended experimental outputs. Alternatively, ‘Fractional Factorial Design—FFD’ can be selected when information about the interactions between media components are not available (usually for a specific strain). Other statistical methods include Plackett–Burman Design (PBD), Taguchi design followed by Central Composite Design (CCD), Box Behnken Design (BBD) and Response Surface Methodology (RSM). The strategy for media optimization could be selected as each method has its advantages and limitations (See Table 1). The utilization of different media components, and physicochemical parameters employed, needs to be evaluated by conducting actual production of the desired metabolites.

Table 1.

Advantages/merits and limitations/demerits of different optimization techniques

Technique	Key feature	Advantages/merits	Limitations/demerits	Reference
One factor at a time (OFAT)	Traditional approach adapted to change only one factor/cause at a time while keeping all other constant	● Easy to interpret results using simple graphs ● Helps to screen critical factors in the absence of previous knowledge ● Commonly used tool for hypothesis testing	● Laborious and time-consuming ● High experimental cost ● Increased number of experiments ● Optimum points can be missed completely ● Major interactions between factors cannot be determined/studied	(Kennedy and Krouse 1999; Abou-Taleb and Galal 2018)
Plackett burman design (PBD)	A two-level screening design for observing the effects of ‘N’ experimental variables in only N + 1 experimental runs	● Minimal number of experimental runs needed ● All factors studied at only 2 levels ● Economical for detecting the main effects ● Only those variables that actually affect the production of desired metabolite are screened ● Screening of non-contributing factors from the contributing factors	● Detection of major effects only with an assumption that there are negligible interactions among the factors ● Applied where factors have additive effects on the output	(Singh et al. 2017; Reddy et al. 1999)
Taguchi method	The experimental design based on an orthogonal array to understand the effect of different parameters on the output of the process by organizing the parameters affecting the process and the levels at which they should be varied	● Minimal experiments to identify the factors that substantially affect the process/product without having to test all possible combinations ● Analyses both the main effects and interactions of two factors ● Quantitative measurement of quality ● Efficient and easier DOE application via implementation of orthogonal arrays ● Quality improvement with reduced costs and time ● Design robust systems by reducing noise factors (uncontrollable variables) that cause performance variability ● Optimal control levels identification using the signal-to-noise ratio (S/N)	● Assumption that high order interactions in the system are negligible ● Complex algorithm with an increased number of calculations ● There is no degree of freedom when all possible factors are considered ● Randomization occurs infrequently ● Relative results ● No exact information on having maximum effect on performance characteristic values ● Relationships between all variables cannot be determined ● Ineffective for a dynamic process (simulation study) ● Applied effectually only at early stages of a process development	(Sinha 2019; Singh et al. 2017; Drożdżyńska et al. 2023; Maghsoodloo et al. 2004; Li et al. 2009)
Factorial designs	The experimental design consisting of two or more factors, each of which has multiple possible values or levels	● This method investigates all possible combinations of a factor's level at every level of every other factors ● Identification of interactions among several variables simultaneously ● Determination of main effects of numerous varying factors ● Analysis of individual and combinatorial effects of factors on dependent variables ● Estimation of the effects of one factor at multiple levels of other factors in a variety of experimental conditions for valid conclusions ● Analysis of effects of one factor at multiple levels of other factors in conjunction with various experimental conditions to draw valid conclusions ● Time saving and economical compared to OFAT	● The number of trials increases with the increase in factors and their levels ● Complexity in experimental setup and statistical analysis increases with analysis of multiple factors at multiple levels ● Uncertainty in accurate interpretation of high-order interactions	(Kennedy and Krouse 1999; Keith 1988; Armstrong et al. 2002)
Response surface methodology (RSM)	Designs that use different mathematical and statistical models and analyses various variables greatly affecting the response of interest	● Efficient application for designing, modelling, prediction and optimisation ● 3D visualisation of data ● Minimal experimental runs yielding comparatively abundant data ● Detection of effects of interactions of the independent parameters, exclusively or in combination with the desired response ● Mathematical models for better process response that are experimentally validated ● Model iteration until the best optimal solution is attained or experimental resources are exhausted	● Erroneous extrapolation beyond the experimental ranges ● Approximation of discrete values in continuous values leads to inaccuracy ● Fitting of experimental data to a second-order polynomial model may not be compatible for all the systems having curvature	(Kennedy and Krouse 1999; Karimifard and Moghaddam 2018)
Artificial neural network (ANN)	Computational model based on machine learning algorithms to mimic learning ability of the human brain	● Ample amount of data can be easily handled (unlimited inputs and outputs ● Enhanced pattern recognition ability ● Training the networks to make decisions based on past learning events ● Mechanistic understanding of system not needed ● Problem representation via Attribute-value pairs (AVP) ● Robust model tolerant to noisy and erroneous training data ● Efficient parallel processing ability ● Capability to learn and model complex and non-linear relationships among data ● No constraints on range of input variables ● Ability to generalize and prediction of unknown data ● Storage of data entirely on network and not on the database allows functioning of network even in the absence of a data pair ● Network shows gradual break-down without an abrupt loss of functioning	● Intolerant of missing data ● Issues of overtraining ● High computational cost ● Problem of data imbalance due to simultaneous optimization by ANN ● Hardware dependent ● Non-reliability due to unexplained functioning of ANN in achieving optimal solutions ● Inability to determine the proper network structure of ANN as it is not structured on specific rules but on trial and errors ● Requires longer training time ● Longer time period needed for convergence of a solution for a dataset with numerous parameters ● Translation of all problems into numerical data only	(Kennedy and Krouse 1999; Caputo et al. 2010; Abdolrasol et al. 2021)
Genetic algorithm (GA)	An adaptive exploratory search algorithm based on principles of genetics and natural selection	● Efficient results for large search space and several parameters ● Steady improvement of solution with time or successive iterations ● Handles noisy data ● Automated selection of new medium ● Efficient parallel processing ability ● Ability to deal with complex problems ● Ability to optimize continuous, discrete and multi-objective functions	● Ineffective application for simple problems ● Insights of the designer not considered ● Difficulty in identification of fitness functions ● Stochastic nature of the algorithm ● Increase in computational complexity with calculation of fitness value after successive generations ● Inappropriate formulation of fitness function, choice of parameters may lead to premature or non-convergence of the algorithm resulting in non-optimal solutions	(Kennedy and Krouse 1999; Banerjee et al. 2021)
Fuzzy logic	Computational approach based on multivalued logic rather than binary logic mimicking the reasoning ability of human brain	● Tolerant of noisy data and missing data	● Needs fine tuning ● Difficulty in devising a set of rules	(Kennedy and Krouse 1999)

Plackett—burman design

This approach screens the most important components required in designing the fermentation media and assists to identify the N + 1 experiment for ‘N variables’. In PBD design, the major factors are taken into consideration and other negligible, non-contributing ones are not considered. Thus, PBD proposes the experimental designs for more than seven factors. For example, when we consider factors such as 7, 11, 15, 19, etc.; the number of experiments would be 8, 12, 16, 20 respectively. For example, if eleven factors are to be screened, then PBD uses 12 runs unlike FFD which requires 16 experimental runs (Goli et al. 2020; Vanaja and Shobha Rani 2007).

Full factorial designs

In the Full Factorial Designs (FuFD) approach, experimental design are structured in such a way that multiple factors can be altered at one instance facilitating the investigation of the cumulative effects of all possible combinations on each other. It aims to run every possible combination to obtain information for individuals and combined effects of factors. For example, if there are four factors such as temperature, pH, microbial strain and inoculum size; each factor with two levels in set up; then FuFD requires 2⁴ = 16 runs. Therefore, if the number of factors and levels increases, then the experiment numbers also increase. Thus, conducting those many experiments then becomes challenging. In that case, partial factorial analysis can be performed using a few combinations (Oimoen 2019). FFD  includes only a 'selected subset' or 'fraction' of the runs. FFD is a logical choice when resources are limited or the number of factors is large where half the number of runs are considered as compared to  FuFD (Antony and Cahyono 2022).

Taguchi method

Full factorial design is difficult or challenging to handle if the number of experiments increases with an increase in the number of process variables. To overcome this difficulty, the orthogonal array-based ‘Taguchi Design’ was developed. This method is based on a highly fractional orthogonal design which simultaneously and independently evaluates two or more parameters (Modi and Desai 2018). Instead of all possible combinations, this method benefits researchers to investigate the involvement of different parameters with few experiments. Pignatiello (1988) provided three features of the Taguchi method such as (i) system strategy, (ii) parameter designing and (iii) tolerance design phase to accomplish the desirable quality of the product. The objective of the Taguchi method is always to maximise the signal-to-noise (S/N) ratio of the response variable. Whereas, RSM methodology is used to find the maximum /minimum or nominal (target) value of the response variable depending upon the objective of the particular experiment.

Box-Behnken design

The Box-Behnken Design (BBD) is an alternative approach to FuFD in which three-levels of each factor are required (Ferreira et al. 2007). The various combinations are either at the midpoints of edges and/or at the centre. BBD uses a second-order polynomial model which precisely defines linear interactions and quadratic effects. In BBD the total number of experiments is determined by the equation, number of experiments = 2 k (k−1) + C (Where k = Number of factors and C = Number of centre points).

Central composite design

The Central Composite Design (CCD) is also used to build a second-order (quadratic) model applicable to RSM. CCD predicts linear and quadratic interaction effects of the medium with the involvement of the factor’s interactions. This design has three sets of experimental runs:

• Factorial design: Two levels are required (high, + 1 and low, − 1).

• Centre points: Consist of experimental runs which have the median values of each factor.

• Star points: Consists of the experimental runs which are matching to the centre points, that considers both below and above the median (Asghar et al. 2014).

BBD differs in CCD as it has fewer design points as compared to the latter one and thus allowing BBD to run the experiments in cheaper ways. Additionally, BBDs have three levels per factor, whereas, CCDs have up to five runs per factor. This subsequently leads to fewer runs in BBDs than CCDs. Since CCDs comprise both factorial and axial points, it is thus more comprehensive than BBDs making it further robust.

Response surface methodology

Response Surface Methodology (RSM) basically represents the collection of computational methods dedicated to estimate the interactions and quadratic effects in a fermentation media (Elfghi and Amin 2014). The RSM is used routinely to determine the best response value, identify improved or optimal process settings, and troubleshoot processes to enhance the quantity of the product under given conditions. In 2021, Ekpenyong et al. discussed the concept of “Path of Steepest Ascent” (PSA), which explains an intermediary gradient protocol in RSM-based optimization experiments. PSA is useful to explore identified significant factors of a process or a system that is close to the optimum through sequential experimentation. Thus, PSA identifies a new operating region for process improvement in substantial ways.

Table 1 Represents the summary of various optimization techniques used in enhancing the production of fungal chitosan along with their advantages and limitations.

Current scenario in the area of fungal chitosan optimization strategies

Optimization studies of FCH were initiated in the early 90 s, where OFAT methods were primitive and popular as compared to the currently available advanced tools and software. Arcidiacono and Kaplan (1992) evaluated the effect of culture and processing protocols on molecular weight distribution of the FCH and biomass from M. rouxii. They reported a maximum yield of FCH (5–10%) from the total biomass. While this study did not focus on optimization of the FCH directly, it did work on optimizing the processing protocol to obtain a high amount of FCH. The study also supported developing optimization strategies like the determination of critical factors affecting the purity and MW of the FCH. Similar studies have been reported by Kim et al. (2001) and Nwe and Stevens (2004) demonstrating the employment of OFAT-based optimization approaches to enhance the yield of FCH for several fermentation parameters including the initial pH conditions. Kim et al. (2001) included aeration, and agitation parameters. Whereas Nwe and Stevens (2004) included varied concentrations of urea to provide substantial evidence of the supportive role of varied parameters in enhancing the yield of FCH. The OFAT approach appears to be implemented frequently in the optimization of FCH. OFAT has been integrated with other methods like BBD, CCD and Taguchi etc. (Habibi et al. 2021, Wang et al. 2008). Instead of focusing only on the nutritional components, the other physical experimental variables can be effectively optimized using method such as BBD, CCD and Taguchi (Abdel-Gawad et al. 2017; Mokhtari-Hosseini et al. 2018; Yonis et al. 2019; Yasrebi et al. 2020). Hence, it can be inferred that by employing low-cost substrate (apple pomace, potato, bean, and corn, etc.), OFAT technique can be well employed as a preliminary step in the optimization strategies and further optimization of the response can be well complemented with techniques such as BBD and CCD that provides a more comprehensive optimization strategy, especially when working with standard media or targeting multiple interacting variables.

Abdel-Gawad et al. (2017) worked on improving FCH yield which involved the use of the BBD at three levels. Employing this statistical method, FCH yield of 7.0% (w/w) was attained from A. niger and the parameters included alkali concentration, incubation temperature and incubation time. Similarly, in 2023, Abo Elsoud et al. used the ANOVA-validated BBD method  and obtained a FCH yield of 3.138 g/l by optimising the nutrient factors like glucose, ammonium sulfate, calcium chloride and physical factors like pH, temperature and incubation time. In 2024, Karamchandani et al. reported an enhanced 2.2 fold of FCH production from C. echinulata using BBD along with PBD significantly. BBD suitability can thus be observed in these studies since multi-factor systems were optimised for enhancing the yield of FCH. Contribution by Abdel-Gawad et al. (2017) and Abo Elsoud et al. (2023) demonstrated that the basic BBD exhibits robust performance in complex systems with statistically proven results. Karamchandani et al. (2024a, b, c) demonstrated that this design (BBD) is flexible enough to adapt with other statistical models to improve its optimization result. Of all the statistical designs, BBD was used quite frequently in these studies since BBD requires fewer runs than necessitated by designs such as CCD. This approach is cost sparing and time saving especially when biological materials are in limited quantities. More importantly, BBD is capable of modelling quadratic interaction among variables which enables for considering an interdependence of the factors such as nutrient concentration, temperature and pH. Further, multiple centre point replicates the possible experimental error resulting in higher statistical reliability and significance of the results. For these reasons, BBD is a method of choice for the enhancement of FCH production in various experimental conditions.

Even though initial reports were not aided by the advanced algorithms provided by software (RSM, Taguchi/Robust design, etc.), they still served as a benchmark for future investigations. In 2004, Goksungur used the CCD approach (one of the designs in RSM) and reported 1.1 g/dm³ of FCH yield from R. oryzae. This noteworthy contribution was one of the pioneers to establish and optimize the concentration of ‘C’ source (sugar) as an influencing parameter when non-standardized components such as molasses (conventional substrate for industrial fermentation) was used as a substrate. In the ten-year span of 2008–2018, several studies dealt with optimizing the production of FCH using an orthogonal array, factorial method, and different designs under the umbrella of RSM technique. Few studies however used more than one design and therefore, some of the studies stood out in a positive way. First, for the initial screening protocols established the moderate levels of the imperative parameters and second in optimizing those parameters. A study by Abo Elsoud et al. (2023) made use of BBD to predict the factors that affect the production of FCH from a selected fungal culture. The characterization of FCH was  based on the properties such as DDA, MW, water-binding capacities (WBC) along with fat-binding capacities (FBC). Collaborators disclosed A. terreus as one of the most potent fungal isolates with maximum productivity (2.92 g/l) in a fermentation medium (g/l) composed of glucose, 35.6; (NH₄)₂SO₄, 4.6; CaCl₂, 0.29; at pH 7.9 and temperature of 23.2 °C up to 10 days. One such study by Vaingankar and Juvekar (2014) used OFAT to determine the moderate or ‘0—level’ for the next optimization phase supported with BBD. For the first time, El Mekawy et al. (2013) and Zhang et al. (2014) used FuFD and PBD followed by BBD and Taguchi / orthogonal array where two optimization methodologies were combined to examine the production of FCH. Here, the first approach (FuFD, PBD) was used for preliminary screening followed by the second (BBD) approach for subsequent optimization of FCH production. This is the first example of such a strategy where the authors employed PBD to identify ideal glucose concentration (g/l), inoculum percentage, the concentration of MgSO₄·7H₂O (g/l) and the time of culture incubation or fermentation period (days) as critical factors for chitin deacetylase (CDA) production by R. japonicus M193. Based on these results, factors were optimized using Taguchi design with an orthogonal array. Optimum conditions of 5 g/l glucose, 5% inoculum level, 0.6 g/l MgSO₄·7H₂O, and 5 days incubation yielded around 2.5% of CT and 547.38 ± 12.06 U/l of CDA. MW of the FCH was found to be 125.63 ± 3.74 kDa having DDA of 78.85 ± 1.68%. A similar approach was explored by El Mekawy et al. (2013) in optimizing FCH yield from A. coerulea where a three-level full factorial design was carried out at the initial stage. The experiments were then continued with the RSM approach (BBD) and demonstrated the highest yield (0.1829 g/l) of FCH (after 72 h, at 32 °C) from A. coerulea when grown in a medium containing 14 g/l ammonium sulphate and 20 g/l glucose.

Another major establishment during the period of 2018–2021, was the use of ‘inoculum size’ as a key parameter. Work by Mokhtari-Hosseini et al. (2018) showed maximum production of 1.08 g/l of FCH from G. lucidum. Research group of Zhang et al. (2014) did use ‘inoculum percent’ as a factor, whereas it was optimized using PBD and Taguchi methods which were known screening methods at that time. Since then, it has became standard to use more than one approach for production of FCH. Habibi et al. (2021) and El-Far (2021) explored the CCD approach and demonstrated its efficiency in optimizing the fermentation conditions for a high yield of FCH. El-Far (2021) reported around 3.1-fold increase in the FCH amount from Penicillium chrysogenum with a high DDA (94%). Varied strains of fungi reported for FCH production have been represented in Fig. 2 and listed in Table 2.

Fig. 2.

A comprehensive summary of diverse fungal species reported for high yield of fungal chitosan (FCH) standardized through optimized parameters over the last three decades (1991–2023). A percentage-wise breakdown for: A Employment of varied optimization techniques—One-factor-at-a-time (OFAT), Box-Behnken Design (BBD), Central Composite Design (CCD), Plackett–Burman Design (PBD), Taguchi, orthogonal array, full factorial, 2² factorial, 2² Central Composite Rotatable Design (CCRD) and Face Centered-Central Composite Design (FC-CCD) to enhance the fungal biomass and FCH yield. B Diverse fungi including Absidia coerulea, Rhizopus oryzae, R. japonicus, Aspergillus niger, A. terreus, A. flavus, Cunninghamella elegans (≈ 16% each) followed by Mucor rouxii (≈ 11%). Other fungi namely, Benjaminiella poitrasii, Penicillium chrysogenum and Trametes versicolor, Gongronella butleri and Ganoderma lucidum (≈ 5% each) explored for FCH production, and C Sunburst illustration denoting the placement and importance of optimization techniques employed for maximizing the FCH production. Reading and interpretation of an image—Different color codes assigned from the inner centre: White circle: Genus, Second pink circle: Species, Third blue circle: Method of Optimization and Outer yellow circle: Optimized parameters

Table 2.

Summary for various optimization techniques employed to standardize several fermentation parameters and accomplish high yield of chitosan from diverse fungal communities

Organism	Method/Design	Parameters	FCH yield	DDA (%)	Media used	Techniques used for characterization	Reference
Aspergillus terreus	OFAT, FC-CCD	• pH, temperature • Ammonium nitrate conc	49 mg/g substrate, 141 mg/g cell weight	88	Apple pomace extract	FTIR, TGA, DSC	(Habibi et al. 2021)
Aspergillus niger	BBD	• Alkali conc • Temperature • Time	7% w/w	83	PDA	Capillary viscometry, UV–Vis, FTIR	(Abdel-Gawad et al. 2017)
Aspergillus flavus	OFAT, CCD	• Chitin conc • Incubation period	27.3 mg/g of substrate	84	PDA	FTIR, Viscometry	(Yonis et al. 2019)
Absidia coerulea	OFAT, Orthogonal array	• Basal substrate • Carbon conc • Nitrogen source	6.12 g/kg	85	Potato, cotton seed hulls, soya bean, corn	GPC-LLS, FTIR, XRD, NMR	(Wang et al. 2008)
Absidia coerulea	Full factorial, BBD	• Carbon conc • Nitrogen conc • Incubation time • Temperature	183 mg/l	74	YPG	FTIR, GPC-SEC	(El Mekawy et al. 2013)
Absidia coerulea	OFAT	• pH • Agitation • Aeration	2.3 g/l	-	Custom optimized	No characterization	(Kim et al. 2001)
Benjaminiella poitrasii	OFAT	• Media composition	51 mg/g of cell weight	94	YPG	FTIR, NMR, Viscometry	(Mane et al. 2017)
Cunninghamella elegans, Rhizopus arrhizus	2 > 2 factorial	• Carbon source • Nitrogen source	49 mg/g	82	CSL, Molasses	IR, Viscometry, XRD, SEM	(Berger et al. 2014)
Cunninghamella elegans	CCRD 2 > 2	• CSL • PPJ conc	37 mg/g of substrate	86	CSL, papaya peel juice	IR, Viscometry, XRD, TGA, DSC	(Berger et al. 2018)
Ganoderma lucidum	CCD	• Inoculum size • Initial pH	1.08 g/l	79	PDB	FE-SEM, XRD, FTIR,	(Mokhtari-Hosseini et al. 2018)
Gongronella butleri	OFAT	• Urea conc • Initial pH	4 g/kg of sweet potato	89	Sweet potato	UV, Gel permeation, HPLC	(New and Stevens 2004)
Mucor rouxii	OFAT	• Incubation period • Agitation, • pH, Temperature	39% of cell weight	78	Date syrup	FTIR, Viscometry	(Chechan et al. 2015)
Mucor rouxii	OFAT	• Extraction protocol	5–10%of biomass	60–90	YPF, BG, TVB, complex media	GPC, IR, UV–Vis, titration	(Arcidiacono and Kaplan 1992)
Penicillium chrysogenum	PBD, CCD	• Culture type • Peptone conc • Incubation period	7.5% (before 4.4%, 3.1 fold increase)	94	Czapek- Dox, SDA, yeast extract	FTIR	(El-Far 2021)
Rhizopus oryzae	CCD	• Aeration • Agitation • Sugar conc	1109 mg/dm3	-	Beet molasses	No characterization, only optimization of FCH	(Göksungur 2004)
Trametes versicolor	Taguchi	• NaOH conc • Time • Temperature	261 mg/l	78	PDA	FTIR	(Yasrebi et al. 2020)

One-Factor-At-a-Time OFAT, Plackett–Burman Design PBD, Central composite design CCD, Face centred central composite design FC-CCD, Central composite rotatable experimental Design CCRD, Corn steep liquor CSL, Potato dextrose broth PDB, Yeast extract peptone glycerol YPG, Sabouraud dextrose agar SDA, Field emission scanning electron microscopy FE-SEM, X-Ray diffraction XRD, Fourier Transform Infrared Spectroscopy FTIR, Differential scanning calorimetry DSC, High-performance liquid chromatography HPLC, gel permeation chromatography, combined with laser light scattering GPC-SEC, Infra-red IR, UV–Vis spectroscopy UV–Visible, Gel permeation chromatography GPC, Nuclear magnetic resonance NMR

Comparison of the literature reported over the last three decades (1991–2023) suggests that fungi namely A. coerulea, C. elegans, R. oryzae. and A. niger have been reported mostly (≈ 16% each; contributing ≈ 64% of the total reports) followed by Mucor rouxii (≈ 11%) and then B. poitrasii, P. chrysogenum and G. butleri and Ga. lucidum and T. versicolor (≈ 5% each) (Fig. 2A). Among different approaches, OFAT was reported frequently (≈ 29%) followed by CCD (≈12%) and OFAT with CCD (≈11%). Other approaches namely BBD, Taguchi, FC-BBD, CCD and factorial design at the individual level were also reported (≈ 6%) (Fig. 2B). Absidia spp. and Cunninghamella spp. are slightly challenging to grow under laboratory conditions in comparison to Rhizopus (Mucoraceae) and Aspergillus (Aspergillaceae). Refer Fig. 2C for ‘Sunburst diagram’ representing the employment of varied optimization techniques in identifying required fermentation parameters for FCH production at the commercial scale.

Design of experiment (DOE)—optimizing media conditions for high yield of fungal chitosan via design expert (stat-ease)

‘OPTIMIZATION’ is defined as the process to make the BEST or rather most effective use of any available resources to achieve desirable results at a successful level. Media components and other growth conditions make a huge difference and a great impact on FCH production. The basic elemental sources should be provided at optimum concentration along with process variables for extraction of maximum quantity of FCH. Conventionally, OFAT approach has been previously used to identify the influence of each factor on fungal growth. OFAT, however, has two basic challenges; that a) it leads to needlessly a greater number of experimental runs and b) the interaction among so many factors become difficult to be analysed. Subsequently, it is challenging to identify the key factor influencing in optimization of media components and process variables through DOE approach. Kaur and Rahi (2024) reported the application of RSM leading to a remarkable improvement in CT yield (from 1.14 g/l to an impressive 4.42 g/l) which is a substantial enhancement (3.34-fold) as compared to unoptimized conditions for A. niger. Additionally, the authors explored the antioxidant activity of A. niger-derived CT, revealing its scavenging activity (32% to 55%) at 1–2 mg/ml. Thus, researchers could optimize the production of CT production from A. niger and also demonstrated its remarkable antioxidant potential, denoting its significance for various applications. Apart from this mushroom, the production of FCH has been accomplished from unexplored fungi (Agaricus bisporus, G. lucidum). Janesch et al. (2020) and Savin et al. (2020) reported the FCH extraction from A. bisporus and G. lucidum, respectively. They added that the viscosity and MW of extracted FCH was comparatively low showing its advantages for diverse applications.

The advancement of information led to the development of software to execute experimental research work in a more precise and easier way. Currently, various software IBM® SPSS® Statistics, Minitab, Design Expert (Stat-ease), JMP, Tableau Server and RStudio are available to optimize microbial products with great success. Preliminary screening of important factors followed by the optimization of variables are mandatory to investigate significant responses. One could design the experiment and subsequently analyse the specific responses which are vital to the fermentation processes. Based on the literature, the following hypothetical example is portrayed to comprehend the experiment planning, designing and understanding the key steps for FCH production. The Design expert software (version 12, Stat-ease) used for analysis and designing the experiment includes the following steps:

1. Listing and selection of all the parameters which could affect the yield of fungal biomass and ultimately increase the FCH production from the intended strains.

2. Screening of parameters that enhances the yield of FCH production.

3. Identification of factors triggering the FCH yield.

4. RSM for FCH optimization using CCD/BBD approaches.

The detailed steps for any experiment design are described as below:

Listing and selection of all the parameters which could affect the yield of fungal biomass and ultimately increasing the FCH production from the intended strains.

The extensive literature review suggests that the basic components for any Zygomycetes fungal growth are carbon source, nitrogen source, incubation time, temperature, pH, agitation speed, and inoculum size.

Screening of parameters that enhances the yield of FCH production

The screening of factors is either performed by fractional factorial, Taguchi design or by PBD. The Taguchi design allows choosing factors ranging between 3 to 21 and thus, assists in characterizing and testing the robustness of the selected parameters. On the other hand, PBD accommodates factors ranging between 4 to 48. The interaction among the factors is supported by mathematical models with linear equations at two levels. The generated Pareto charts and half-normal plots which are statistically significant suggesting the main factors involved. Various parameters are studied on the basis of their low (−1), medium (0) and high values (+ 1). The statistical design is represented in terms of coded factors. By default, the high level of any particular factor is coded as + 1, whereas the low level is coded as −1. The coded equation is useful for identifying the relative impact of the factors by comparing their factor coefficients. For example, if literature suggests that the carbon concentration for any particular fungus is 20 g/l, then the low, medium and high values could be set as below 20 g/l (quantified as ‘−1 level’), at 20 g/l (quantified as ‘0 level’) and above 20 g/l (quantified as ‘ + 1 level’). Depending upon the number of factors, experiments can be designed through PBD or Taguchi method.

The work published by Habibi et al. (2021); Vaingankar and Juvekar (2014) and Göksungur (2004) showed that the above factors are helpful in attaining high amounts of biomass as well as FCH. As an example, the design matrix for screening of six factors is illustrated in Table S1 (Supplementary material). The Design Expert software simplifies the experimental design and the initial shake flask experiments support in achieving the response for each variable. The responses produced due to the combination of factors support in predicting a few important factors accountable for that particular response. The ‘ANOVA results’ help in analysing the model whether it is ‘SIGNIFICANT’ or ‘INSIGNIFICANT’. It is determined by the ‘P-values’. If the value is < 0.0500, it indicates model terms are significant. The Design Expert software predicts the coefficient of determination (R² value). The equation in terms of coded factors can be used to make predictions about the responses for given levels of each factor. The half-normal plots and Pareto charts are helpful in identifying the dynamic factors generated based on the statistical significance of the factors and the responses.

Identification of factors triggering the FCH yield:

The screening of the significant factors obtained by the above method generates the ‘Pareto chart’ to represent the screened parameters. These parameters are then analysed for obtaining a high FCH yield.

Response surface methodology for FCH optimization using CCD/BBD approaches

The RSM is used to optimize the identified factors responsible for FCH production. The identified factors are then used in the BBD/CCD model. The response surface plots generated illustrate the dependency of three screened factors and their interaction with each other to maximize the yield of FCH (Fig. 3).

Fig. 3.

Sequential illustration of the methodology to be followed for large-scale production of chitosan (FCH) from Zygomycetes fungi through statistical  Design of Experiments (DOE). The process begins with (1) Defining the objective (optimizing the production of biopolymer, followed by (2) Screening the factors (media components, experimental conditions) influencing FCH yield using appropriate screening designs (Plackett–Burman Design or Taguchi design). (3) The output response, aligned with the objective is selected (FCH yield). (4) Identifying factor levels (5) For example—low, medium, high level for each factor followed by 6) The development of Response Surface Methodology (RSM) based design including Box-Behnken Design (BBD) and Central Composite Design (CCD). (6) Conducting the shake flask experiments based on the generated experimental design, and (7) Conduct of the experiment at shake flask level. (8) Collection of data and analysis, graphical representation (8a: A Box-Behnken Design; 8b: The 3-dimensional surface plot ; 8c: Contour plot), validation and confirmation of models to draw meaningful interpretations from the model for large-scale production of FCH

Commercial/industrial prospects of fungal chitosan production and optimization

The production and extraction of FCH from the fungal cell wall through fermentation processes is feasible economically and a green synthesis approach. FCH has many applications in food and beverage, cosmetics, biomedical and tissue engineering (Ghosh and Urban 2009; Khan et al. 2019, 2020; Duan et al. 2019; Wang et al. 2019; Tang et al. 2020). Among varied industrial sectors, a wide variety of research focuses on the biomedical application of FCH due to its molecular characteristics. For example, FCH is soluble in wide physiological pH ranges due to its poly-cationic nature and lower antigen effect (Khan et al. 2020). FCH also has considerable LMW (≈ 20–30 kDa) which is quite a desirable property (Huq et al. 2022). As a result, the demand for FCH is becoming widespread and the use of microbial production requires effective large-scale industrial fermentation technology.

The large-scale fermentation cost including capital investment and manufacturing operations are quite high. The market aspect requires various alternatives to reduce cost through exploring different microbial sources, alternative substrates and methods of production (Huq et al. 2022). According to Allied Market research (Chitosan 2023), the total market value of CH was $1.7 billion in 2019 including all the sources of FCH extraction. The market value is projected to reach $4.7 billion in 2027 at a Compound annual growth rate (CAGR) of 14.5% (https://www.alliedmarketresearch.com/chitosan-market). The report includes the names of major key players in the FCH market such as Advanced Biopolymers AS (Norway), Heppe Medical Chitosan GmbH (Germany), G.T.C. Bio Corporation (China), Primex EHF (Iceland), Kitozyme S.A. (Belgium), Novamatrix (Norway), Agratech International, Inc. (U.S.), Golden-Shell Pharmaceutical Co. Ltd. (China), Qingdao Yunzhou Biochemistry Co. Ltd. (China), and Panvo Organics Private Ltd (India). Commercial CH cost ranges from USD $ 22 per Kg (India) to USD $ 237 per 100 gm (Sigma Aldrich, USA). This large variation in the costing is due to the downstream processing used to isolate purified products.

The properties of FCH are ultimately better as its downstream processing is less polluting and laborious. The growth of fungi under optimized conditions favour higher biomass production and subsequently higher FCH yields. Additionally, the use of cheap, renewable, or waste products can be a better alternative for the growth of fungi providing economic feasibility over the use of expensive dehydrated or complex standard media. Habibi et al. (2021) used apple pomace and additionally added ammonium nitrate (5.0 g/l concentration) to produce FCH with a biomass of 140.9 mg/cell grown on substrate yield of 0.387 g/cell/g/substrate. Other physiological parameters such as temperature (30 °C) and initial pH 5.98 were also optimized. Scientists have also used agro-industrial wastes such as CSL and cassava wastewater for FCH production from C. elegans. The appropriate use of DOE along with cheap substrate is helpful for taking down the cost of the FCH production process.

Life cycle assessment (LCA) helps to make a verdict for an optimum usage of various resources for obtaining the targeted product. In view of this, Muñoz et al. (2018) discussed that 33 kg fresh weight of shrimp shell is required for the production of 1 kg of CT. The shrimp cells undergo demineralization (using dilute HCl) and deproteinization (using NaOH) to manufacture 1 kg of CT. Additionally, the CCH production from CT requires other resources such as NaOH, water, electricity, and fuel. The production of 1 kg CCH requires 1.4 kg of CT and the process of deacetylation again requires the use of NaOH, 250 L of water and 1.06 kWh of electricity or wood fuel (31 MJ). The outlet of CCH production includes ‘C’, ‘N’ sources of wastewater and salts. The produce is ultimately used as ‘N’ fertilizer for agricultural fields, while the remaining salts from chemicals can be disposed of in landfills. There is no literature available for the LCA of FCH production. Accepted practice suggests that a fermentation-oriented production will make the overall process more sustainable in commercial settings. However, LCA is needed to evaluate other aspects like the energy consumption of bioreactor, downstream processing, waste management, etc. A thorough investigation is thus required to assess the sustainability of CCH and FCH production. Figure 4 highlights the selection of suitable optimization approaches that can be employed for achieving high biomass along with a substantial yield of FCH from different fungi.

Fig. 4.

Decision tree for selecting applicable optimization approaches for enhancing the production of fungal chitosan (FCH) to address the issues compared to unoptimized conditions. The process begins with identifying the potent fungi capable of producing FCH followed by the identification of factors influencing the yield. (1) Choosing appropriate screening techniques to select contributing variables. (2) Establishing varied experimental ranges—low, medium, and high levels. (3) Application of optimization tools to maximize the response i.e. enhancing the yield of FCH. (4) Data collection, analysis and validation of data for large-scale production of the biopolymer

Concluding remarks and future prospects

The supply of CCH relies mainly on crustacean and shellfish waste as a major source of extraction. FCH sources however are believed to be better alternatives considering their superior qualities. We recommend a few measures that could be adopted to enhance the “two-step extraction procedures (acid and base treatment)” in respect to the extraction of FCH and CCH to improve yield, efficiency and quality. Firstly, it is mandatory to optimize the Reaction Parameters where optimal concentrations of Acid and Base are essential to confirm operational demineralization and deproteinization processes and minimize the degradation of the biopolymers. Likewise optimal temperatures and time is also required during the two-step processes to enhance reaction kinetics without degrading CH. It is essential to get rid-of the impurities during washing and drying of the crude product. Monitoring DDA through analytical techniques can authenticate the quality of the CH. By considering the above measures, the “two-step extraction process” for the extraction of CH can be made proficient, economically viable to produce high-quality biopolymers for innumerable applications. Factors such as environmental pollution and efforts to develop green chemistry have sparked a strong interest in natural-source materials. One of the undeniable advantages of materials derived from natural sources is their availability. This can sometimes result in lower costs than those obtained through chemical synthesis. Nonetheless, their practical application is primarily determined by specific parameters that a given material must meet to be suitable for a particular application. The production process of the fungal glycans could be utilized as a substrate for the growth of mycelia which will ultimately aid in a circular economy. The cell wall of Zygomycetes fungi possess higher percentages of CT and CH and therefore is an ideal candidate for extraction of those valuable biomaterials in a cost-effective manner. Typical dehydrated or complex media are popular for fungal growth and therefore have been utilized for FCH production. Initially, listing as well as the selection of all the parameters followed by screening and identification of factors promoting biomass are mandatory. The employment of an appropriate method for designing the experiment adds on to the commercial-level production of FCH. DOE tools such as RSM, PBD, CCD, BBD, Taguchi approach have a crucial role in improving fungal biomass along with the yield of FCH. The Taguchi method aims to maximise the signal-to-noise (S/N) ratio of the response variable. Whereas, RSM methodology is used to find the maximum /minimum or nominal (target) value of the response variable depending upon the objective of the particular experiment. The competent use of optimization techniques, either individually or in combination supports high production of FCH. Use of appropriate strains with a low number of experiments could be less laborious. Some of the challenges such as the high fermentation costs in FCH fermentation could be reduced by discovering renewable cheap or low-cost waste substrates. Chatterjee et al. (2023) used the collagenous fleshing waste of leather processing industry for the growth and production of FCH from Zygomycetes fungi. Use of low-cost substrates thereby leads to a reduction in investment costs. Priyanka et al. (2023) reported use of banana peels as cost-effective substrate for FCH production. Thus, the use of inexpensive and agricultural waste residues or substrates could be helpful in resolving the economic issues associated with myco-based FCH industry and directing the commercial-scale production of the biopolymers.

Current advances in statistical, AI tools, software and mathematical models are reducing the challenges associated with microbial technologies in an economical way as compared to conventional approaches such as OFAT. With the increasing advent of computer-based technologies, experiments could be designed using Face centered central composite design (FCCCD). The ANN could be employed for analysing, predicting and validating FCH biosynthesis from potent fungal strains. Additionally, the complexity in experimental setup analysis, especially multiple factors at multiple levels, can be achieved with the help of statistical models and improve the yield of FCH. The substitution of conventional processes to myco-based approach may necessitate the evaluation of economic and ecological factors. Looking at the current FCH production scenario, LCA becomes an important aspect. High-risk environmental challenges associated with the conventional extraction process must be considered for a greener future. This technology should be adopted for optimum use of all the resources in the future. The adoption of dependable software and statistical tools would undoubtedly empower researchers to statistically design their experiments and accomplish optimization processes for microbial metabolites with great success. Considering future applications, careful design of target-specific proteins could be investigated to improve FCH synthesis in fungal cell walls. The use of inducible promoters with varying strengths will aid in modulating effector expression levels and optimizing the performance of fungi at the genetic level. Using AI and bioinformatics tools to investigate protein–protein interactions can help to develop more effective production of enzymes like chitin synthase and deacetylase, leading to higher FCH production.

Supplementary Information

Below is the link to the electronic supplementary material.

Funding

Authors from the Department of Microbiology are grateful to Savitribai Phule Pune University, Pune, Maharashtra, India for Financial support.

Data avilability

The raw data supporting the conclusion of this article will be made available by the authors upon reasonable request.

Declarations

Conflict of interest

The authors report no conflict of interest.