Computational Models of Algae Metabolism for Industrial Applications

Abstract

Many algae species exhibit phenotypes that are of great interest for industrial applications. Systems biology approaches including large-scale metabolic reconstructions can guide and enhance production of biobased commodities such as fuel and human therapeutic proteins. Here we review the existing computational models of metabolism in algal species, as well as the databases and tools that are available to improve the sustainability of algae-to-energy production at industrial scales. As this field has progressed with genome sequencing, computational techniques, novel annotation tools and databases, systems biology will continue to evolve to advance the appeal of algae to a wide array of industries.

Introduction

Microalgae fossil records extend back more than three billion years, and over that period algae have become a vital and ubiquitous component of the Earth's biosphere.¹ Their importance is primarily due to their high photosynthetic productivity, which creates the foundation for many ecosystems.² With more than 40,000 algal species currently identified, microalgae have evolved to succeed in most environments on earth.³ Many microalgal species are autotrophic, meaning they convert sunlight and CO₂ into biomass, and include types of green algae (Chlorophyta), yellow-green algae (Xanthophyta), golden algae (Chrysophyta), red algae (Rhodophyta), diatoms, and dinoflagellates.⁴ This diversity of species contributes to much of the current interest in using microalgae for a variety of commercial purposes, ranging from nutraceutical applications to biofuel production. Here we focus our discussion on eukaryotic species of algae and exclude cyanobacteria, or blue-green prokaryotic algae, due to their distinct cellular properties.

Industrial production strategies have generally sought to capitalize or improve production of one or more of the following outputs from algae: long-chained polyunsaturated fatty acids, food colorants, animal feed and supplementation, cosmetics, hydrogen, wastewater treatment, CO₂ fixation, human therapeutic proteins, and biofuel production.^{5

–13} Thus, algal species have the potential to become “cellular factories” upon genetic re-engineering to improve production and yield of desired commodities.

Despite the promise of commercial applications for microalgae, algal biotechnology remains in its infancy. Terrestrial crops and animals have been bred over thousands of years and selected for the creation of high-yield species for agricultural needs. It stands to reason that efforts to cultivate algae on a large scale will not be based solely on wild-type species but will require some degree of genetic modification before algae can make economically viable contributions to any sector.¹⁴ However, our current limited understanding of how to perform genetic engineering of a species for the creation of target strains with economically favorable characteristics has hindered the industrialization of algal biotechnology.

Academic and commercial interest in the properties of algae metabolism has increased substantially in recent years (Figs. 1A-B). Advances in genome sequencing have allowed for the sequencing of at least 20 microalgal genomes, including the smallest free-living eukaryote Ostreococcus tauri, the unicellular green algae Chlamydomonas reinhardtii, and the planktonic species Botryococcus braunii, with applications across a range of industries (Fig. 1C).^{9,15

–76} Key species of interest include B. braunii, C. reinhardtii, Dunaliella salina, and Phaedactylum tricornutum, although to date only a select number of engineered strains and products have been commercialized.

Fig. 1.

Snapshot of algae species with sequenced genomes in research and industrial applications. Number of publications cited in PubMed per year from 2000 through 2012 (A). Search terms included “algae and metabolism”; “Chlamydomonas and metabolism”; and “Ostreococcus and metabolism.” Number of patents issued in the United States per year with algae or alga in title or abstract, according to the United States Patent and Trademark Office (www.uspto.gov) (B). For sequenced algal genomes, the number of publications for the indicated application cited on PubMed (C). References for the indicated species are listed as follows: B. braunii, ^{18

–31} C. reinhardtii, ^{9,32

–51} D. salina, ^52

–58 Micromonaspusilla CCMP1545,⁵⁹ O. lucimarinus,⁶⁰ O. tauri, ^60,61 Volvox carteri,^60,62 Phaeodactylum tricornutum,^63

–69 Thalassiosira pseudonana,^70,71 Chondrus crispus,⁷² Cyanidioschyzon merolae,^60,73 and Galdieria sulphuraria.^74
–76

Furthermore, within the broader field of systems biology, advances in whole genome sequencing have paved the way for the reconstruction of organism-specific metabolic networks, with genome-scale networks having been created for several prokaryotes as well as for an increasing number of more complex, eukaryotic species.⁷⁷ Metabolic network models are systems-level, genomic computational tools that can assist in the reengineering of key metabolic pathways for industrial-level production of desired products such as lipids and biofuels.⁷⁸ Additional applications include bioremediation, wastewater treatment, and carbon dioxide (CO₂) fixation. For example, a genome-scale metabolic network for the model eukaryote Saccharomyces cerevisiae was probed using in silico engineering strategies to increase ethanol yields on glucose and production of fumaric acid, a potential petroleum substitute.^79,80 Similarly, metabolic network reconstructions of algae hold promise to aid in the identification of genetic modifications for the creation of optimized strains with commercial potential for desired commodity production. To date, metabolic network reconstructions have been generated for four different algal species: C. reinhardtii, Ostreococcus lucimarinus, O. tauri, and B. braunii.

In this review, we highlight the current state of algae metabolic network modeling and optimization tools with an emphasis on the application of in silico metabolic engineering reconstructions. We use this information to provide an analysis of how systems-based approaches can inform and be informed by large-scale supply chain constraints in the algae-to-energy landscape, and how this might influence commercial development strategies.

Metabolic Engineering Reconstruction Process

Metabolic network reconstructions are mathematically organized structures assembled from various data sources including biochemical, genomic, and cell phenotypic data. Assembled networks account for known reactions, enzymes, and genes that are part of the metabolic pathways of an organism. Generating a metabolic network reconstruction is an iterative process involving several rounds of manual curation and experimental validation. The first step in creating a genome-scale metabolic reconstruction is to obtain the most recent version available of the target organism's sequenced genome and annotation to identify open reading frames (ORFs) and assign them functions by comparing them to genes associated with proteins of known functions using available online tools such as Uniprot and ChlamyCyc.^81,82

Once these molecular functions have been assigned, reactions are then associated with the annotated genes, taking into account stoichiometry and reversibility, using databases like KEGG and PredAlgo.^83
–85 Additionally, to create the most biologically accurate model with accurate network connectivity, cellular localization must be taken into account, especially for complex eukaryotic organisms such as algae. Unlike the other steps of the reconstruction process, which rely heavily on the comparison of the functionality of genes across different species and organisms, compartmentation of reactions requires organism-specific knowledge; however, there is often little to no evidence for reaction localization.⁸⁶ Thus, in the event that the localization of a particular reaction is unknown, researchers typically infer the compartment based on related reactions in homologous species, relying on transporter reactions to carry metabolites from one compartment to another. In the case of the well-studied plant species Arabidopsis thaliana and its reconstructed network AraGEM, compartmentation of reactions was completed manually based on the literature, and in many cases, reactions were localized by default to the cytosol to avoid relying on uncharacterized transport reactions.^86,87 Certainly, as biological knowledge increases, so too will the quality of available network reconstructions and associated reaction localization.

There are several freely available tools that semi-automate the draft reconstruction process, such as Pathway Tools and Model SEED; however, extensive manual curation is still required.^88
–90 Table 1 highlights key resources publicly available for the reconstruction of metabolic networks, including several algae-specific annotation tools.^{81,82,85,88

–104} Ultimately, this extensive and iterative manual curation process results in an organism-specific, genome-scale metabolic reconstruction that summarizes the metabolic pathways of a target organism by documenting information such as the gene-protein-reaction (GPR) relationships, reaction stoichiometry, reaction reversibility, and cellular compartment. We provide an overview of the process to generate an in silico metabolic network for algae commodity production, including computational tools for optimized strain selection, in Fig. 2.

Fig. 2.

Snapshot of the in silico metabolic engineering process of algal strains for industrial commodity production. Genome sequencing allows for network formation and mathematical representation by a stoichiometric matrix. Computational tools may aid in the selection of engineered strains for large industrial-scale commodity production.

Table 1.

Genome and Pathway Resources Available for Metabolic Reconstructions

RESOURCES	APPLICATION	LINK
GENOME RESOURCES
Algal Functional Annotation Tool⁹¹	Algae-specific genome annotation tool	http://pathways.mcdb.ucla.edu
Chlamydomonas Connection	Chlamydomonas genome database	www.chlamy.org
TAIR^92,93	Arabidopsis genome database	www.arabidopsis.org
PROTEIN RESOURCES
BRENDA^94,95	Enzyme classification database	www.brenda-enzymes.org
ExPASy⁹⁶	Genomic and proteomic database	www.expasy.org
PredAlgo⁸⁵	Protein intracellular localization tool	giavap-genomes.ibpc.fr/predalgo
UniProt^82,97	Protein sequence database	www.uniprot.org
PATHWAY RESOURCES
AraCyc⁹⁸	Arabidopsis metabolic pathway database	www.arabidopsis.org/tools/aracyc
BioCyc⁸¹	Organism specific pathway database	http://biocyc.org
ChlamyCyc⁸¹	Chlamydomonas metabolic pathway database	http://pmn.plantcyc.org
KEGG¹⁰⁰	Gene, protein, reaction and pathway database	www.genome.jp/kegg
MetaCyc^101,102	Metabolic pathway database	http://metacyc.org
Reactome¹⁰³	Biological pathway database	www.reactome.org
RECONSTRUCTION RESOURCES
BiGG¹⁰⁴	Metabolic network reconstruction database	http://bigg.ucsd.edu
Model SEED⁸⁹	Draft metabolic network reconstruction tool	http://theseed.org
Pathway Tools⁹⁰	Draft metabolic network reconstruction tool	http://pathwaytools.org

In Silico Methods for Strain Optimization

Once the metabolic network reconstruction has been manually curated, constraint-based modeling approaches can then be performed to characterize aspects of the organism's metabolism. One commonly used approach, flux balance analysis (FBA), can be used to predict the fluxes of metabolites through the metabolic network when optimized toward a biologically relevant objective, such as cell growth or the production of a metabolite of interest. FBA simulations can be performed using software tools like the COBRA Toolbox, a freely available software package that runs in MATLAB, and the BioMet Toolbox, an online resource for integrating high-throughput data for systems analysis.^105,106

To perform FBA, the metabolic network reconstruction must first be converted into a mathematical framework in the form of a stoichiometric matrix, S, where the rows of S correspond to the metabolites present in the reconstruction and the columns of S correspond to the reactions present in the reconstruction (Fig. 2). As such, each element of the matrix denotes the stoichiometric coefficient of a particular metabolite in a given reaction. Subsequently, a system of linear equations is established by multiplying the stoichiometric matrix by a column vector, v, which accounts for the fluxes through the reactions in the system, and setting this product equal to the rate of change of concentrations of all the species in the system. This vector of concentrations is equal to zero under steady-state conditions. The assumption that the metabolic network operates at steady-state is well justified under many conditions of interest.¹⁰⁷

Because the resulting system of equations is underdetermined (ie, there are typically more reactions than metabolites in a reconstruction), FBA converts this system to a linear programming problem by optimizing for a particular flux called the objective function. Examples of objective functions include biomass production as an estimate for cell growth yield or the production of a metabolite of commercial interest. Additional constraints, such as enzyme capacities, maximum uptake, secretion rates and thermodynamic constraints, can be placed on the system to further narrow the solution space of feasible flux distributions.¹⁰⁸ Using these constraints and the specified objective function, FBA then calculates a distribution of fluxes through the metabolic network corresponding to the optimal value of the objective.

Since its introduction, several extensions have been made to FBA that have potential applications for in silico metabolic engineering. For example, in silico gene knockout simulations can be performed by taking advantage of the GPR relationship defined in the reconstruction to predict the effect that a model gene deletion or a combination of gene deletions will have on either biomass production or on the yield of a metabolite of interest.¹⁰⁹ However, given the size of genome-scale metabolic network reconstructions, the amount of data generated from in silico gene deletion combination simulations has proven to be challenging to manage.¹¹⁰ To combat the immense volume of combinatorial data, several frameworks have been proposed to guide in silico gene knockout studies in a systematic and biologically relevant manner. OptKnock, for instance, is structured as a two-part optimization problem that evaluates the effect of gene knockouts to maximize both the overproduction of a metabolite of interest as well as the biomass.¹¹¹ Several extensions to OptKnock such as OptGene, OptStrain, OptORF, and OptReg have been proposed to help further guide in silico metabolic engineering gene deletion simulations.^112

–115 OptReg, for instance, extends OptKnock by allowing for gene up- and/or down-regulation, enabling the development of a broader range of potential metabolic engineering strategies.

Additionally, several FBA-based algorithms have been developed to depict more clearly the objective function that best characterizes the metabolism of mutant organisms. Studies have suggested that, unlike the wild-type phenotype, mutant organisms that have not been exposed to long-term evolutionary pressure do not operate in an optimized manner in regards to biomass production. Rather, in response to a genetic perturbation, the mutant tries to limit the difference in flux distribution between itself and the wild-type phenotype. In an effort to capture this altered objective function, algorithms like Minimization Of Metabolic Adjustment (MOMA) and Regulatory On/Off Minimization (ROOM) have been developed to minimize the predicted redistribution of metabolic fluxes in mutant phenotypes after a gene knockout by applying a minimal distance metric that limits the overall change in flux distribution compared to the wild type.^116,117

Further aiding in strain design, several methods like flux variability analysis (FVA) and flux coupling analysis (FCA) have been developed to study the structural and topological properties of metabolic networks. FVA identifies the range of possible fluxes for each reaction in a network given a certain objective function value.¹¹⁸ FCA identifies pairs of metabolic fluxes in the network that are coupled together directionally, partially, or fully, as well as blocked reactions that are unable to carry a flux under certain conditions.^118
–120

Finally, genome-scale metabolic network reconstructions are an excellent platform for the integration of high-throughput data, allowing for the creation of context-specific computational models that have the potential to aid in identifying mutant strains suitable for the overproduction of a metabolite of commercial interest. As such, several algorithms have been developed for the integration of a variety of “omics” data, with special interest given to the integration of expression data due to the wealth of transcriptomic data currently available. One such method, Gene Inactivity Moderated by Metabolism and Expression (GIMME), employs a user-defined threshold to enable the methodical removal of reactions below the threshold.¹²¹ Additional methods include the Integrative Metabolic Analysis Tool (IMAT), which creates a functioning model for fluxes based on mRNA transcript levels without relying on a user-specified objective function, and Metabolic Adjustment by Differential Expression (MADE), which uses the statistical significance of changes in gene expression measurements to turn model genes on and off.^122–123 Existing methods and integration techniques are improving continually. The current state of such methods, including E-flux and Probabilistic Regulation of Metabolism (PROM), has been reviewed by Blazier et al.¹²⁴

The use of FBA, FVA, and in silico gene knockout simulations performed on algal metabolic network reconstructions is discussed in further detail in the following section. Other algorithms, such as the computationally efficient combinatorial gene knockout algorithms, and the methods for omics data integration, have yet to be used with algae reconstructions, primarily because genome-scale algae metabolic network reconstruction development and analysis is still in its infancy.¹²⁵ Nevertheless, these algorithms have much potential to aid metabolic engineering strategies in genome-scale metabolic network reconstructions and provide biological insight into algal species.

Algal Metabolic Network Reconstructions

Although the genomes of at least 20 different algal species have been sequenced (Fig. 1C; http://genome.jgi-psf.org), metabolic network modeling of algae remains a relatively young field. As metabolic networks involve functional annotation of all known metabolic reactions, organisms with fully sequenced genomes remain the best candidates for comprehensive network reconstructions. To date, genome-scale metabolic network reconstructions have been generated for only three algal species: C. reinhardtii, O. lucimarinus, and O. tauri. Additionally, the metabolic pathways involved in liquid hydrogen production of B. braunii were reconstructed using transcriptomic data in the absence of a fully sequenced genome.¹⁷ A summary of the existing metabolic reconstructions for these species is shown in Table 2, along with the corresponding number of metabolites and reactions accounted for in the models.^{17,126

–131}

Table 2.

Sequenced Genomes and Existing Metabolic Models for Algae^a

SPECIES	DESCRIPTION	GENOME SIZE (MB)	MODEL RECONSTRUCTIONS	NUMBER OF COMPARTMENTS	METABOLITES; REACTIONS
Botryococcus braunii ¹⁷	Green, colonial fresh water algae	166	Yes	1	N/A
Chlamydomonas reinhardtii ^126 –130	Unicellular, soil-dwelling green algae	121	Primary	3	458; 484
			iAM303	9	259; 467
			Thermo	1	278; 280
			iRC1080	10	1,068; 2,090
			AlgaGEM	4	1,869; 1,725
Ostreococcus lucimarinus ¹³¹	Small, marine phytoplankton	13.2	Yes	1	1,100; 964
Ostreococcus tauri ¹³¹	Smallest eukaryote	12.6	Yes	1	1,014; 871

Compartments for each reconstruction are listed as follows: B. braunii was not compartmentalized; primary network of C. reinhardtii includes cytosol, mitochondria, and chloroplast; iAM303 includes cytoplasm, endoplasmic reticulum, extracellular, Golgi apparatus, lysosome, mitochondria, nucleus, peroxisome, and plasma membrane; C. reinhardtii thermodynamic model is not compartmentalized; iRC1080 includes mitochondria, glyoxysome, extracellular space, nucleus, Golgi apparatus, flagellum, eyespot, thylakoid, chloroplast, and cytosol; AlgaGEM is divided into cytosol, mitochondria, plastid, and microbody (peroxisome). Both O. lucimarinus and O. tauri models were not compartmentalized.

Prior to reconstructing the full genome-scale metabolic network of C. reinhardtii, several initial efforts focused on the reconstruction of the central metabolic network, with each successive iteration accounting for additional properties of the algal species. The first published model for algae focused on the central metabolism of C. reinhardtii and included 458 intracellular metabolites and 484 metabolic reactions, accounting for fatty acid, amino acid, and nucleotide synthesis as well as glycolysis and the citric acid cycle.¹²⁶ This model accounted for reaction and metabolite localization in three compartments: the cytosol, the mitochondria, and the chloroplast. Using flux balance analysis, the intracellular metabolic fluxes were predicted for C. reinhardtii growth under three different conditions: autotrophic, heterotrophic, and mixotrophic. Each of these three growth conditions was modeled by modifying the limits on CO₂ and acetate uptake reactions as well as setting a constraint for the amount of light absorbed by the system.

Using an iterative methodology that combined bioinformatics and experimental techniques, a second central metabolic network reconstruction, iAM303, was generated with 259 reactions and 467 metabolites across several cellular compartments including the cytosol, mitochondria, chloroplast, glyoxysome, and flagellum.¹²⁷ In order to validate the model, the results from FBA after optimization of the network for either biomass or adenosine triphosphate (ATP) production were compared to experimental values found in the literature. Additionally, in silico gene-knockout simulations and FVA were used to propose genetic engineering strategies for increased hydrogen production. A non-compartmentalized model of the primary metabolism was also created by incorporating thermodynamic parameters with constraint-based modeling techniques to determine the extent of light-driven respiration.¹²⁸ By constructing a two-step objective function, including both the growth rate of cells and the photon uptake rate, this study allowed for simulation of biomass growth under low light conditions, revealing interactions between respiratory activity and photosynthesis in C. reinhardtii.

More recently, two genome-scale metabolic network reconstructions have been generated for C. reinhardtii. One such reconstruction, iRC1080, accounts for the function of 1,068 metabolites and 2,190 reactions localized across 10 compartments, and was the first metabolic network reconstruction to account for photon absorption and growth simulations quantitatively under various light sources.¹²⁹ Using the C. reinhardtii central metabolism reconstruction iAM303 as a starting point, Chang et al. added reactions on a pathway-by-pathway basis using more than 250 publications. iRC1080 also significantly expanded the number of lipid metabolic pathways over preceding metabolic reconstructions, enabling the application of in silico metabolic engineering strategies aimed at enhancing lipid production. In order to validate the model, in silico gene knockout simulations were performed and subsequently compared to published data on mutant C. reinhardtii phenotypes. FBA and FVA were subsequently used to simulate adaptations in metabolic response to changes in light influx for contribution of metabolic pathways to biomass production, including lipid synthesis and protein production.¹³²

AlgaGEM, another C. reinhardtii genome-scale metabolic network reconstruction, accounts for the function of 1,869 metabolites and 1,725 metabolic reactions across several cellular compartments such as the cytoplasm, the mitochondrion, the plastid, and the microbody. Compartmentalization was determined from published literature or by comparing homologs to a metabolic model of the plant species A. thaliana. To identify gaps in the model, FBA was used to evaluate the model's ability to produce major biomass components under autotrophic, heterotrophic, and mixotrophic conditions. AlgaGEM was subsequently used to predict metabolic targets for the increased production of H₂ by performing in silico gene knockout simulations on the network.¹³⁰

In addition to those curated for C. reinhardtii, genome-scale metabolic network reconstructions have been generated for the green algae O. lucimarinus and O. tauri. Ostreococcus species are prevalent as microalgae and are ideal organisms to study because of their simplicity and phylogenetic relationship as early-diverging green plants.¹³¹ Furthermore, O. tauri is becoming popular as a model organism for metabolic studies as it has the smallest eukaryotic genome and an unusually strong adaptation to nutrient stress, demonstrated in part by its ability to respond to low nitrogen conditions—suggesting a fast global metabolic response.¹³³ To generate genome-scale metabolic network reconstructions for both O. lucimarinus and O. tauri, draft networks were first reconstructed for each species using the KEGG database. Subsequently, a sophisticated FBA-based gap-filling algorithm that took into account phylogenetic distance among species was applied to the two draft reconstructions, resulting in more complete and accurate networks. Additionally, this algorithm highlighted gaps in the annotation of genes for both Ostreococcus species.

Finally, in the absence of a fully sequenced genome, a model for the organism B. braunii, a high oil-producing algal strain, was constructed using transcriptomic data and various databases including KEGG, MetaCyc, and Reactome.¹⁷ The ability of B. braunii to produce large volumes of terpenoid liquid hydrocarbons, which are similar to crude fossil fuels, motivated the study of the genes and pathways involved in terpene production. Using a large-scale next-generation sequence read dataset, the Department of Energy Joint Genome Institute produced a metabolic network through various annotation methods, including the B. braunii Showa web-based annotation tool, FrameDP, and the NetStart 1.0 Server, to identify genes and pathways key for the production of terpene compounds and precursors. As a result of this study, an annotated transcriptomic database for B. braunii was made publicly available online, paving the way for future in silico metabolic engineering analyses such as gene knockout simulations to optimize the production of terpenoid liquid hydrocarbons.

Simulations for the majority of C. reinhardtii models, including iAM303, iRC1080, and AlgaGEM, were completed with the COBRA toolbox implemented in MATLAB, freely available from the openCOBRA project (http://opencobra.sourceforge.net/).¹⁰⁵ Each of the Ostreococcus models, once converted to SBML format, were then processed with the COBRA toolbox v2.0.¹³⁴

Industrial Applications of Biotechnology

A key to making algae-to-energy systems viable and sustainable in the long run will be to tie emerging metabolic network models with the large-scale systems drivers that will determine the industrial relevance of these processes. Ultimately, the technical and economic viability of a scaled up algae-based biotechnology sector, especially for products to be produced in large quantities such as fuels, will depend largely on the availability of land, water, nutrients, and CO₂, as well as markets for byproducts, among other constraints.¹³⁵ A great majority of existing algae biotechnology research has taken for granted the supply chains and large-scale systems that will need to be put in place to support such an industry at relevant scales.¹³⁶ This is starting to change as life cycle analysis and other tools from the emerging field of industrial ecology are being used to try and understand more about ancillary processes.¹³⁷ By evaluating biological viability and industrial sustainability in unison, in silico metabolic engineering and industrial ecology can be used to optimize parameters in a way that maximizes the potential of industrial algae processes.

A number of companies currently produce compounds from algae including ethanol, hydrogen food colorants, and biofuel precursors.¹³⁸ Several chemical companies, such as Synthetic Genomics (La Jolla, CA) and Sapphire Energy (San Diego, CA), have employed metabolic engineering strategies to improve fuel yields from algae.¹³⁹ Solazyme (San Francisco, CA) has produced and begun commercializing biobased fuels from heterotrophic algae, although precise synthetic biology and metabolic engineering strategies have remained proprietary.¹⁴⁰ Other algae production companies and their methods for biofuel conversion have been previously described.¹⁴¹ A large number of startup biotechnology companies are also working to leverage computational metabolic methods in the production of cosmetics, pharmaceuticals, and specialty chemicals; Rosetta Green (Rehovot, Israel), for example, is working toward the successful production of human growth hormone from algae.

Conclusions

Efforts to deploy metabolic optimization strategies are limited primarily by the complexity inherent in understanding biological systems. To overcome this complexity, efforts to leverage metabolic network models have tended to focus on single factor optimization. In the context of algae-biofuel production, for example, there has been a heavy focus on the maximization of lipid yields.¹⁴ Empirical data from years of laboratory work suggest that very high lipid yields (>50% mass) are possible, but that these yields also result in lower growth rates.¹⁴² If the total lipid productivity is higher, even though the cell growth rate is lower, then such efforts to maximize lipid production would be worthwhile. While tradeoff analysis is relatively simple on a pathway-by-pathway basis, it becomes challenging for increasingly complex metabolic systems. Metabolic network reconstructions along with in silico analyses such as FVA can be used to aid in the tradeoff assessment for both simple and complex metabolic systems.

A great opportunity exists to couple metabolic engineering with regional materials flow analysis. Material flow analysis is a quantitative tool for measuring material stocks and flows at the scale of regions or nations.¹⁴³ Continuing on with the algae-biofuel example, this coupling would enable algae farmers to select strains that are both regionally appropriate and capable of achieving maximum profits under current market conditions. Knowing how water composition and average temperatures affect growth rates in a variety of strains could help an algae farmer select a family of algae that would be the most viable strains for their site. The developer could additionally refine strain selection by considering which byproducts would be most valuable under prevailing market conditions. Obtaining this information using traditional laboratory approaches would be immensely expensive and limited because it would not allow for the manipulation of algae cells for some generally desirable traits; however, as genome-scale metabolic network reconstructions become more encompassing, they may be used to address such concerns as optimal strain selection.

Interestingly, the computational structure of most tools in industrial ecology, including life cycle assessment and material flow analysis, has parallels with metabolic network reconstructions. These models are generally comprised of large systems of equations that describe material flows resulting from deliberate human activity in the economy or via indirect or undesirable pathways in the form of pollution.¹⁴⁴ These systems of equations can be manipulated depending on the system boundaries or functional unit of interest.¹⁴⁵ Over the past decade, methods for obtaining data that are temporally and spatially specific for material and energy flows have resulted in several large commercial databases that greatly facilitate the development of life cycle and material flow models.¹⁴⁶ The similarities between the numerical methods used in both problems suggest that they could one day be integrated to yield algae strain selection tools that span huge spatial scales.

There are certain phenotypes that will be desirable for all industrially relevant algae. Most importantly, species that maximize available CO₂ will be prized given how difficult it will be to set up a CO₂ supply chain to deliver CO₂ to large-scale algae cultivation facilities. Most analyses have tended to assume that CO₂ will be obtained from a large point source, such as a coal or natural gas-burning power plant.¹⁴⁷ Coal-fired plants are ubiquitous and their flue gas contains high levels (∼12%) of CO₂. The flue gas tends to be dirty, however, containing significant particulate matter such as SO₂ and NOx, which can impact the pH of pond water if it is not remediated first. Flue gas from natural gas power plants is much cleaner but it has lower levels of CO₂ and, consequently, the costs of separating out the CO₂ are higher. Given how involved it will likely be to supply large industrial facilities with CO₂, designing algae to use CO₂ more efficiently would improve the efficiency and bottom line of the company.¹⁴⁸ In silico gene knockout simulations could be performed on metabolic network reconstructions using algorithms like OptKnock and MOMA to aid in the design process of mutant algae that more efficiently metabolize CO₂.

Other important phenotypic characteristics include the algae species' ability to tolerate high salinity growth media; produce compounds that would enable them to outcompete invasive species; remain neutrally buoyant in growth ponds; or be capable of autoflocculation.¹⁴⁹ These characteristics, which can be understood in part in the context of metabolic network models, will have important ramifications on the engineering processes necessary to use algae for biofuels production. Employing these metabolic network models will help ensure that genetic modifications are being carried out in the most rational and efficient way possible, and that the most appropriate tools are being developed for the burgeoning industry.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Menetrez

. An overview of algae biofuel production and potential environmental impact. Environ Sci Technol, 2012; 46:7073–7085.

Cardinale

, Matulich

, Hooper

et al. The functional role of producer diversity in ecosystems. Am J Bot, 2011; 98:572–592.

, Sommerfeld

, Jarvis

et al. Microalgal triacylglycerols as feedstocks for biofuel production: Perspectives and advances. Plant J, 2008; 54:621–639.

Gong

, Jiang

. Biodiesel production with microalgae as feedstock: From strains to biodiesel. Biotechnol Lett, 2011; 33:1269–1284.

Apt

, Behrens

. Commercial developments in microalgal biotechnology. J Phycol, 1999; 35:215–226.

Spolaore

, Joannis-Cassan

, Duran

, Isambert

. Commercial applications of microalgae. J Biosci Bioeng, 2006; 101:87–96.

Brown

, Jeffrey

, Volkman

, Dunstan

. Nutritional properties of microalgae for mariculture. Aquaculture, 1997; 151:315–331.

Pulz

, Gross

. Valuable products from biotechnology of microalgae. Appl Microbiol Biotechnol, 2004; 65:635–648.

Skjanes

, Rebours

, Lindblad

. Potential for green microalgae to produce hydrogen, pharmaceuticals and other high value products in a combined process. Crit Rev Biotechnol, 2013; 33,2:172–215.

10.

Dalrymple

, Halfhide

, Udom

et al. Wastewater use in algae production for generation of renewable resources: A review and preliminary results. Aquat Biosyst, 2013; 9,1:2.

11.

Kishimoto

, Okakura

, Nagashimaet

et al. Co₂ fixation and oil production using microalgae. J Ferment Bioeng, 1994; 78:479–482.

12.

Franklin

, Mayfield

. Recent developments in the production of human therapeutic proteins in eukaryotic algae. Expert Opin Biol Ther, 2005; 5:225–235.

13.

Wang

, Wang

, Zhang

, Meldrum

. Application of synthetic biology in cyanobacteria and algae. Front Microbiol, 2012; 3:344.

14.

Georgianna

, Mayfield

. Exploiting diversity and synthetic biology for the production of algal biofuels. Nature, 2012; 488:329–335.

15.

Derelle

, Ferraz

, Rombauts

et al. Genome analysis of the smallest free-living eukaryote Ostreococcus tauri unveils many unique features. Proc Natl Acad Sci, 2006; 103:11647–11652.

16.

Merchant

, Prochnik

, Vallon

et al. The Chlamydomonas genome reveals the evolution of key animal and plant functions. Science, 2007; 318:245–250.

17.

Molnar

, Lopez

, Wisecaver

et al.

Bio-crude transcriptomics: Gene discovery and metabolic network reconstruction for the biosynthesis of the terpenome of the hydrocarbon oil-producing green alga, Botryococcus braunii race B (Showa)

BMC Genomics, 2012; 13:576.

18.

Boyd

, Champagne

, McGinn

et al. Switchable hydrophilicity solvents for lipid extraction from microalgae for biofuel production. Bioresour Technol, 2012; 118:628–632.

19.

Kim

, Ramanan

, Cho

et al. Simple, rapid and cost-effective method for high quality nucleic acids extraction from different strains of Botryococcus braunii. PloS One, 2012; 7:e37770.

20.

Zhang

, Cheng

, Gao

et al. Mechanism of lipid extraction from Botryococcus braunii FACHB 357 in a biphasic bioreactor. J Biotechnol, 2011; 154:281–284.

21.

Rao

, Sarada

, Baskaran

, Ravishankar

. Antioxidant activity of Botryococcus braunii extract elucidated in vitro models. J Agric Food Chem, 2006; 54:4593–4599.

22.

Buono

, Langellotti

, Martello

et al. Biological activities of dermatological interest by the water extract of the microalga Botryococcus braunii. Arch Dermatol Res, 2012; 304:755–764.

23.

Fang

, Chiu

, Wu

et al. Fatty acids in Botryococcus braunii accelerate topical delivery of flurbiprofen into and across skin. Int J Pharm, 2004; 276:163–173.

24.

Honda

, Boonnorat

, Chiemchaisri

et al. Carbon dioxide capture and nutrients removal utilizing treated sewage by concentrated microalgae cultivation in a membrane photobioreactor. Bioresour Technol, 2012; 125:59–64.

25.

, Liu

, Tian

. Growth characteristics of Botryococcus braunii 765 under high CO₂ concentration in photobioreactor. Bioresour Technol, 2011; 102:130–134.

26.

Sydney

, Sturm

, de Carvalho

et al. Potential carbon dioxide fixation by industrially important microalgae. Bioresour Technol, 2010; 101:5892–5896.

27.

Yoo

, Jun

, Lee

et al. Selection of microalgae for lipid production under high levels carbon dioxide. Bioresour Technol, 2010; 101,Suppl 1:S71–S74.

28.

Casadevall

, Dif

, Largeau

et al. Studies on batch and continuous cultures of Botryococcus braunii: Hydrocarbon production in relation to physiological state, cell ultrastructure, and phosphate nutrition. Biotechnol Bioeng, 1985; 27:286–295.

29.

Ranga Rao

, Ravishankar

, Sarada

. Cultivation of green alga Botryococcus braunii in raceway, circular ponds under outdoor conditions and its growth, hydrocarbon production. Bioresour Technol, 2012; 123:528–533.

30.

Magota

, Saga

, Okada

, Atobe

, Imou

. Effect of thermal pretreatments on hydrocarbon recovery from Botryococcus braunii. Bioresour Technol, 2012; 123:195–198.

31.

Song

, Qin

, Su

et al.

Micronutrient requirements for growth and hydrocarbon production in the oil producing green alga Botryococcus braunii (Chlorophyta)

PloS One, 2012; 7:e41459.

32.

La Russa

, Bogen

, Uhmeyer

et al. Functional analysis of three type-2 DGAT homologue genes for triacylglycerol production in the green microalga Chlamydomonas reinhardtii. J Biotechnol, 2012; 162:13–20.

33.

Castruita

, Casero

, Karpowicz

et al. Systems biology approach in Chlamydomonas reveals connections between copper nutrition and multiple metabolic steps. Plant Cell, 2011; 23:1273–1292.

34.

Vila

, Galvan

, Fernandez

, Leon

. Ketocarotenoid biosynthesis in transgenic microalgae expressing a foreign beta-C-4-carotene oxygenase gene. Methods Mol Biol, 2012; 892:283–295.

35.

Merchant

, Allen

, Kropat

et al. Between a rock and a hard place: Trace element nutrition in Chlamydomonas. Biochim Biophys Acta, 2006; 1763:578–594.

36.

Yoon

, Kim

, Li

et al. Transgenic microalgae expressing Escherichia coli AppA phytase as feed additive to reduce phytate excretion in the manure of young broiler chicks. Appl Microbiol Biotechnol, 2011; 91:553–563.

37.

Genkov

, Meyer

, Griffiths

, Spreitzer

. Functional hybrid rubisco enzymes with plant small subunits and algal large subunits: Engineered rbcS cDNA for expression in Chlamydomonas. J Biological Chem, 2010; 285:19833–19841.

38.

Timmins

, Thomas-Hall

, Darling

et al. Phylogenetic and molecular analysis of hydrogen-producing green algae. J Exp Bot, 2009; 60:1691–1702.

39.

Burgess

, Tamburic

, Zemichael

et al. Solar-driven hydrogen production in green algae. Adv Appl Microbiol, 2011; 75:71–110.

40.

Philipps

, Krawietz

, Hemschemeier

, Happe

. A pyruvate formate lyase-deficient Chlamydomonas reinhardtii strain provides evidence for a link between fermentation and hydrogen production in green algae. Plant J, 2011; 66:330–340.

41.

Baba

, Suzuki

, Shiraiwa

. Proteomic analysis of high-CO(2)-inducible extracellular proteins in the unicellular green alga, Chlamydomonas reinhardtii. Plant Cell Physiol, 2011; 52:1302–1314.

42.

Kliphuis

, Martens

, Janssen

, Wijffels

. Effect of O(2):CO(2) ratio on the primary metabolism of Chlamydomonas reinhardtii. Biotechnol Bioeng, 2011, 10.1002/bit.23194

43.

Kong

, Li

, Martinez

, Chen

, Ruan

. Culture of microalgae Chlamydomonas reinhardtii in wastewater for biomass feedstock production. Appl Biochem Biotechnol, 2010; 160:9–18.

44.

Wang

, Duanmu

, Spalding

. Carbon dioxide concentrating mechanism in Chlamydomonas reinhardtii: Inorganic carbon transport and CO₂ recapture. Photosynth Res, 2011; 109:115–122.

45.

Ohnishi

, Mukherjee

, Tsujikawa

et al. Expression of a low CO(2)-inducible protein, LCI1, increases inorganic carbon uptake in the green alga Chlamydomonas reinhardtii. Plant Cell, 2010; 22:3105–3117.

46.

Renberg

, Johansson

, Shutova

et al. A metabolomic approach to study major metabolite changes during acclimation to limiting CO₂ in Chlamydomonas reinhardtii. Plant Physiol, 2010; 154:187–196.

47.

Tran

, Van

, Barrera

et al. Production of unique immunotoxin cancer therapeutics in algal chloroplasts. Proc Natl Acad Sci, 2013; 110:E15–22.

48.

Rasala

, Mayfield

. The microalga Chlamydomonas reinhardtii as a platform for the production of human protein therapeutics. Bioeng Bugs, 2011; 2:50–54.

49.

Rasala

, Muto

, Sullivan

, Mayfield

. Improved heterologous protein expression in the chloroplast of Chlamydomonas reinhardtii through promoter and 5′ untranslated region optimization. Plant Biotechnol J, 2011; 9:674–683.

50.

James

, Hocart

, Hillier

et al. Temperature modulation of fatty acid profiles for biofuel production in nitrogen deprived Chlamydomonas reinhardtii. Bioresour Technol, 2013; 127:441–447.

51.

Cakmak

, Angun

, Demiray

et al. Differential effects of nitrogen and sulfur deprivation on growth and biodiesel feedstock production of Chlamydomonas reinhardtii. Biotechnol Bioeng, 2012; 109:1947–1957.

52.

Lamers

, Janssen

, De Vos

et al. Carotenoid and fatty acid metabolism in nitrogen-starved Dunaliella salina, a unicellular green microalga. J Biotechnol, 2012; 162:21–27.

53.

Kleinegris

, Janssen

, Brandenburg

, Wijffels

. Continuous production of carotenoids from Dunaliella salina. Enzyme Microb Technol, 2011; 48:253–259.

54.

Prieto

, Pedro Canavate

, Garcia-Gonzalez

. Assessment of carotenoid production by Dunaliella salina in different culture systems and operation regimes. J Biotechnol, 2011; 151:180–185.

55.

Yang

, Jia

, Su

et al. Utilization of CO₂ and biomass char derived from pyrolysis of Dunaliella salina: The effects of steam and catalyst on CO and H₂ gas production. Bioresour Technol, 2012; 110:676–681.

56.

Hosseini Tafreshi

, Shariati

. Dunaliella biotechnology: Methods and applications. J Appl Microbiol, 2009; 107:14–35.

57.

Taylor

, Rand

, Caldwell

. Treatment with algae extracts promotes flocculation, and enhances growth and neutral lipid content in Nannochloropsis oculata—a candidate for biofuel production. Mar Biotechnol, 2012; 14:774–781.

58.

Davidi

, Katz

, Pick

. Characterization of major lipid droplet proteins from Dunaliella. Planta, 2012; 236:19–33.

59.

Petrie

, Shrestha

, Mansour

et al. Metabolic engineering of omega-3 long-chain polyunsaturated fatty acids in plants using an acyl-CoA Delta6-desaturase with omega3-preference from the marine microalga Micromonas pusilla. Metab Eng, 2010; 12:233–240.

60.

Misra

, Panda

, Parida

, Mishra

. Phylogenomic study of lipid genes involved in microalgal biofuel production-candidate gene mining and metabolic pathway analyses. Evol Bioinform Online, 2012; 8:545–564.

61.

Sayanova

, Ruiz-Lopez

, Haslam

, Napier

. The role of Delta6-desaturase acyl-carrier specificity in the efficient synthesis of long-chain polyunsaturated fatty acids in transgenic plants. Plant Biotechnol J, 2012; 10:195–206.

62.

Yamano

, Fujita

, Fukuzawa

. Photosynthetic characteristics of a multicellular green alga Volvox carteri in response to external CO₂ levels possibly regulated by CCM1/CIA5 ortholog. Photosynthe Res, 2011; 109:151–159.

63.

Valenzuela

, Mazurie

, Carlson

et al. Potential role of multiple carbon fixation pathways during lipid accumulation in Phaeodactylum tricornutum. Biotechnol Biofuels, 2012; 5:40.

64.

Ryckebosch

, Muylaert

, Eeckhout

et al. Influence of drying and storage on lipid and carotenoid stability of the microalga Phaeodactylum tricornutum. J Agric Food Chem, 2011; 59:11063–11069.

65.

Hempel

, Lau

, Klingl

, Maier

. Algae as protein factories: Expression of a human antibody and the respective antigen in the diatom Phaeodactylum tricornutum. PloS One, 2011; 6:e28424.

66.

Desbois

, Lebl

, Yan

, Smith

. Isolation and structural characterisation of two antibacterial free fatty acids from the marine diatom, Phaeodactylum tricornutum. Appl Microbiol Biotechnol, 2008; 81:755–764.

67.

Desbois

, Mearns-Spragg

, Smith

. A fatty acid from the diatom Phaeodactylum tricornutum is antibacterial against diverse bacteria including multi-resistant Staphylococcus aureus (MRSA) Mar Biotechnol, 2099; 11:45–52.

68.

Horst

, Parker

, Dennis

et al. Treatment of Phaeodactylum tricornutum cells with papain facilitates lipid extraction. J Biotechnol, 2012; 162:40–49.

69.

Wileman

, Ozkan

, Berberoglu

. Rheological properties of algae slurries for minimizing harvesting energy requirements in biofuel production. Bioresour Technol, 2012; 104:432–439.

70.

Zendejas

, Benke

, Lane

et al.

Characterization of the acylglycerols and resulting biodiesel derived from vegetable oil and microalgae (Thalassiosira pseudonana and Phaeodactylum tricornutum)

Biotechnol Bioeng, 2012; 109:1146–1154.

71.

Jiang

, Yoshida

, Quigg

. Photosynthetic performance, lipid production and biomass composition in response to nitrogen limitation in marine microalgae. Plant Physiol Biochem, 2012; 54:70–77.

72.

van Ginneken

, Helsper

, de Visser

et al. Polyunsaturated fatty acids in various macroalgal species from North Atlantic and tropical seas. Lipids Health Dis, 2011; 10:104.

73.

Sato

, Moriyama

. Genomic and biochemical analysis of lipid biosynthesis in the unicellular rhodophyte Cyanidioschyzon merolae: lack of a plastidic desaturation pathway results in the coupled pathway of galactolipid synthesis. Eukaryot Cell, 2007; 6:1006–1017.

74.

Graziani

, Schiavo

, Nicolai

et al. Microalgae as human food: Chemical and nutritional characteristics of the thermo-acidophilic microalga Galdieria sulphuraria. Food Funct, 2013; 4:144–152.

75.

Eriksen

. Production of phycocyanin—a pigment with applications in biology, biotechnology, foods and medicine. Appl Microbiol Biotechnol, 2008; 80:1–14.

76.

Weber

, Oesterhelt

, Gross

et al. EST-analysis of the thermo-acidophilic red microalga Galdieria sulphuraria reveals potential for lipid A biosynthesis and unveils the pathway of carbon export from rhodoplasts. Plant Mol Biol, 2004; 55:17–32.

77.

Kim

, Sohn

, Lee

. Metabolic network modeling and simulation for drug targeting and discovery. Biotechnol J, 2012; 7:330–342.

78.

Feist

, Palsson

. The biomass objective function. Curr Opin Microbiol, 201; 13:344–349.

79.

Bro

, Regenberg

, Forster

, Nielsen

. In silico aided metabolic engineering of Saccharomyces cerevisiae for improved bioethanol production. Metab Eng, 2006; 8:102–111.

80.

, Zou

, Chen

et al. Fumaric acid production in Saccharomyces cerevisiae by in silico aided metabolic engineering. PloS One, 2012; 7:e52086.

81.

May

, Christian

, Kempa

, Walther

. ChlamyCyc: An integrative systems biology database and web-portal for Chlamydomonas reinhardtii. BMC Genomics, 2009; 10:209.

82.

UniProt Consortium. The Universal Protein Resource (UniProt) in 2010. Nucleic Acids Res, 2010; 38:D142–148.

83.

Kanehisa

, Araki

, Goto

et al. KEGG for linking genomes to life and the environment. Nucleic Acids Res, 2008; 36:D480–484.

84.

Chang

, Scheer

, Grote

et al. BRENDA, AMENDA and FRENDA the enzyme information system: New content and tools in 2009. Nucleic Acids Res, 2009; 37:D588–592.

85.

Tardif

, Atteia

, Specht

et al. PredAlgo: A new subcellular localization prediction tool dedicated to green algae. Mol Bio Evol, 2012; 29:3625–3639.

86.

de Oliveira Dal'Molin

, Nielsen

. Plant genome-scale metabolic reconstruction and modelling. Curr Opin Biotechnol, 2013; 24:271–277.

87.

de Oliveira Dal'Molin

, Quek

, Palfreyman

et al. AraGEM, a genome-scale reconstruction of the primary metabolic network in Arabidopsis. Plant Physiol, 2010; 152:579–589.

88.

Pinney

, Shirley

, McConkey

, Westhead

. metaSHARK: Software for automated metabolic network prediction from DNA sequence and its application to the genomes of Plasmodium falciparum and Eimeria tenella. Nucleic Acids Res, 2005; 33:1399–1409.

89.

Henry

, DeJongh

, Best

et al. High-throughput generation, optimization and analysis of genome-scale metabolic models. Nat Biotechnol, 2010; 28:977–982.

90.

Karp

, Paley

, Romero

. The Pathway Tools software. Bioinformatics, 2002; 18,Suppl 1:S225–232.

91.

Lopez

, Casero

, Cokus

et al. Algal Functional Annotation Tool: A web-based analysis suite to functionally interpret large gene lists using integrated annotation and expression data. BMC Bioinformatics, 2011; 12:282.

92.

Swarbreck

, Wilks

, Lamesch

et al. The Arabidopsis Information Resource (TAIR): Gene structure and function annotation. Nucleic Acids Res, 2008; 36:D1009–1014.

93.

Lamesch

, Berardini

, Li

et al. The Arabidopsis Information Resource (TAIR): Improved gene annotation and new tools. Nucleic Acids Res, 2012; 40:D1202–1210.

94.

Schomburg

, Chang

, Hofmann

et al. BRENDA: A resource for enzyme data and metabolic information. Trends Biochemic Sci, 2002; 27:54–56.

95.

Schomburg

, Chang

, Placzek

et al. BRENDA in 2013: Integrated reactions, kinetic data, enzyme function data, improved disease classification: New options and contents in BRENDA. Nucleic Acids Res, 2013; 41:D764–772.

96.

Artimo

, Jonnalagedda

, Arnold

et al. ExPASy: SIB bioinformatics resource portal. Nucleic Acids Res, 2012; 40:W597–603.

97.

Magrane

, Consortium

. UniProt Knowledgebase: A hub of integrated protein data. Database (Oxford), 2011; bar009.

98.

Mueller

, Zhang

, Rhee

. AraCyc: A biochemical pathway database for Arabidopsis. Plant Physiol, 2003; 132:453–460.

99.

Karp

, Ouzounis

, Moore-Kochlacs

et al. Expansion of the BioCyc collection of pathway/genome databases to 160 genomes. Nucleic Acids Res, 2005; 33:6083–6089.

100.

Kanehisa

, Goto

. KEGG: Kyoto encyclopedia of genes and genomes. Nucleic Acids Res, 2000; 28:27–30.

101.

Caspi

, Foerster

, Fulcher

et al. MetaCyc: A multiorganism database of metabolic pathways and enzymes. Nucleic Acids Res, 2006; 34:D511–516.

102.

Caspi

, Altman

, Dale

et al. The MetaCyc database of metabolic pathways and enzymes and the BioCyc collection of pathway/genome databases. Nucleic Acids Res, 2010; 38:D473–479.

103.

Joshi-Tope

, Gillespie

, Vastrik

et al. Reactome: A knowledgebase of biological pathways. Nucleic Acids Res, 2005; 33:D428–D432.

104.

Schellenberger

, Park

, Conrad

, Palsson

. BiGG: A biochemical genetic and genomic knowledgebase of large scale metabolic reconstructions. BMC Bioinformatics, 2010; 11:213.

105.

Becker

, Feist

, Mo

et al. Quantitative prediction of cellular metabolism with constraint-based models: the COBRA Toolbox. Nat Protoc, 2007; 2:727–738.

106.

Cvijovic

, Olivares-Hernández

, Agren

et al. BioMet Toolbox: Genome-wide analysis of metabolism. Nucleic Acids Res, 2010; 38:W144–149.

107.

Orth

, Thiele

, Palsson

. What is flux balance analysis? Nat Biotechnol, 2010; 28:245–248.

108.

Price

, Reed

, Palsson

. Genome-scale models of microbial cells: Evaluating the consequences of constraints. Nat Rev Microbiol, 2004; 2:886–897.

109.

Park

, Kim

, Lee

. Constraints-based genome-scale metabolic simulation for systems metabolic engineering. Biotechnol Adv, 2009; 27:979–988.

110.

Patil

, Akesson

, Nielsen

. Use of genome-scale microbial models for metabolic engineering. Curr Opin Biotechnol, 2004; 15:64–69.

111.

Burgard

, Pharkya

, Maranas

. Optknock: A bilevel programming framework for identifying gene knockout strategies for microbial strain optimization. Biotechnol Bioeng, 2003; 84:647–657.

112.

Patil

, Rocha

, Forster

, Nielsen

. Evolutionary programming as a platform for in silico metabolic engineering. BMC Bioinformatics, 2005; 6:308.

113.

Pharkya

, Burgard

, Maranas

. OptStrain: A computational framework for redesign of microbial production systems. Genome Res, 2004; 14:2367–2376.

114.

Pharkya

, Maranas

. An optimization framework for identifying reaction activation/inhibition or elimination candidates for overproduction in microbial systems. Metab Eng, 2006; 8:1–13.

115.

Kim

, Reed

. OptORF: Optimal metabolic and regulatory perturbations for metabolic engineering of microbial strains. BMC Syst Biol, 2010; 4:53.

116.

Segre

, Vitkup

, Church

. Analysis of optimality in natural and perturbed metabolic networks. Proc Natl Acad Sci, 2002; 99:15112–15117.

117.

Shlomi

, Berkman

, Ruppin

. Regulatory on/off minimization of metabolic flux changes after genetic perturbations. Proc Natl Acad Sci, 2005; 102:7695–7700.

118.

Gudmundsson

, Thiele

. Computationally efficient flux variability analysis. BMC Bioinformatics, 2010; 11:489.

119.

Mahadevan

, Schilling

. The effects of alternate optimal solutions in constraint-based genome-scale metabolic models. Metab Eng, 2033; 5:264–276.

120.

Burgard

, Nikolaev

, Schilling

, Maranas

. Flux coupling analysis of genome-scale metabolic network reconstructions. Genome Res, 2004; 14:301–312.

121.

Becker

, Palsson

. Context-specific metabolic networks are consistent with experiments. PLoS Comput Biol, 2008; 4:e1000082.

122.

Zur

, Ruppin

, Shlomi

. iMAT: An integrative metabolic analysis tool. Bioinformatics, 2010; 26:3140–3142.

123.

Jensen

, Papin

. Functional integration of a metabolic network model and expression data without arbitrary thresholding. Bioinformatics, 2011; 27:541–547.

124.

Blazier

, Papin

. Integration of expression data in genome-scale metabolic network reconstructions. Front Physiol, 2012; 3:299.

125.

Ren

, Zeng

, Qian

. Adaptive bi-level programming for optimal gene knockouts for targeted overproduction under phenotypic constraints. BMC Bioinformatics, 2013; 14,Suppl 2:S17.

126.

Boyle

, Morgan

. Flux balance analysis of primary metabolism in Chlamydomonas reinhardtii. BMC Syst Biol, 2009; 3:4.

127.

Manichaikul

, Ghamsari

, Hom

et al. Metabolic network analysis integrated with transcript verification for sequenced genomes. Nat Methods, 2009; 6:589–592.

128.

Cogne

, Rügen

, Bockmayr

et al. A model-based method for investigating bioenergetic processes in autotrophically growing eukaryotic microalgae: Application to the green algae Chlamydomonas reinhardtii. Biotechnol Prog, 2011; 27:631–640.

129.

Chang

, Ghamsari

, Manichaikul

et al. Metabolic network reconstruction of Chlamydomonas offers insight into light-driven algal metabolism. Mol Syst Biol, 2011; 7:518.

130.

Dal'Molin

, Quek

, Palfreyman

, Nielsen

. AlgaGEM—a genome-scale metabolic reconstruction of algae based on the Chlamydomonas reinhardtii genome. BMC Genomics, 2011; 12,Suppl 4:S5.

131.

Krumholz

, Yang

, Weisenhorn

et al. Genome-wide metabolic network reconstruction of the picoalga Ostreococcus. J Exp Bot, 2012; 63:2353–2362.

132.

Rugen

, Bockmayr

, Legrand

, Cogne

. Network reduction in metabolic pathway analysis: Elucidation of the key pathways involved in the photoautotrophic growth of the green alga Chlamydomonas reinhardtii. Metab Eng, 2012; 14:458–467.

133.

Le Bihan

, Martin

, Chirnside

et al. Shotgun proteomic analysis of the unicellular alga Ostreococcus tauri. J Proteomics, 2011; 74:2060–2070.

134.

Schellenberger

, Que

, Fleming

et al. Quantitative prediction of cellular metabolism with constraint-based models: The COBRA Toolbox v2.0. Nature Protoc, 2011; 6:1290–1307.

135.

Clarens

, Resurreccion

, White

, Colosi

. Environmental life cycle comparison of algae to other bioenergy feedstocks. Environ Sci Technol, 2010; 44:1813–1819.

136.

Liu

, Clarens

, Colosi

. Algae biodiesel has potential despite inconclusive results to date. Bioresour Technol, 2012; 104:803–806.

137.

Resurreccion

, Colosi

, White

, Clarens

. Comparison of algae cultivation methods for bioenergy production using a combined life cycle assessment and life cycle costing approach. Bioresour Technol, 2012; 126:298–306.

138.

Schmidt

, Lin-Schmidt

, Chamberlin

et al. Metabolic systems analysis to advance algal biotechnology. Biotechnol J, 2010; 5:660–670.

139.

Waltz

. Biotech's green gold? Nature Biotechnol, 2009; 27:15–18.

140.

Singh

. Facts, growth, and opportunities in industrial biotechnology. Org Process Res Dev, 2001; 175–179.

141.

Singh

, Cu

. Commercialization potential of microalgae for biofuels production. Renew Sust Energ Rev, 2010; 14:2596–2610.

142.

Thomas

, Seibert

DLR

, Alden

et al. Yields, photosynthetic efficiencies and proximate composition of dense marine microalgal cultures. II. Dunaliella primolecta and Tetraselmis suecica experiments. Biomass, 1984; 5:211–225.

143.

Hawkins

, Hendrickson

, Higgins

et al. A mixed-unit input-output model for environmental life-cycle assessment and material flow analysis. Environ Sci Technol, 2007; 41:1024–1031.

144.

Hendrickson

, Horvath

, Joshi

, Lave

. Economic input-output models for environmental life-cycle assessment. Environ Sci Technol, 1998; 32:184A–191A.

145.

Heijungs

, Suh

. The Computational Structure of Life Cycle Assessment. Boston, MA: Kluwer Academic Publishers, Dordrecht, 2002.

146.

Weidema

. ecoinvent data v2.0. Vol. 2008. Duebendorf, Switzerland: Swiss Federal Laboratories for Materials Testing and Research, 2007.

147.

Campbell

, Beer

, Batten

. Life cycle assessment of biodiesel production from microalgae in ponds. Bioresour Technol, 2010; 102:50–56.

148.

Khoo

, Tan

RBH

. Life cycle evaluation of CO₂ recovery and mineral sequestration alternatives. Environ Prog, 2066; 25:208–217.

149.

Stephenson

, Kazamia

, Dennis

et al. Life-cycle assessment of potential algal biodiesel production in the United Kingdom: A comparison of raceways and air-lift tubular bioreactors. Energy Fuels, 2010; 24:4062–4077.