Bioaugmentation enhances dark fermentative hydrogen production in cultures exposed to short-term temperature fluctuations.

Hydrogen-producing mixed cultures were subjected to a 48-h downward or upward temperature fluctuation from 55 to 35 or 75 °C. Hydrogen production was monitored during the fluctuations and for three consecutive batch cultivations at 55 °C to evaluate the impact of temperature fluctuations and bioaugmentation with synthetic mixed culture of known H2 producers either during or after the fluctuation. Without augmentation, H2 production was significantly reduced during the downward temperature fluctuation and no H2 was produced during the upward fluctuation. H2 production improved significantly during temperature fluctuation when bioaugmentation was applied to cultures exposed to downward or upward temperatures. However, when bioaugmentation was applied after the fluctuation, i.e., when the cultures were returned to 55 °C, the H2 yields obtained were between 1.6 and 5% higher than when bioaugmentation was applied during the fluctuation. Thus, the results indicate the usefulness of bioaugmentation in process recovery, especially if bioaugmentation time is optimised.

Conversion of methane to organic acids is a widely found trait among gammaproteobacterial methanotrophs of freshwater lake and pond ecosystems.

IMPORTANCE: Aerobic gammaproteobacterial methanotrophic bacteria (gMOB) play an important role in reducing methane emissions from freshwater ecosystems. In hypoxic conditions prevalent near oxic-anoxic interfaces, gMOB potentially shift their metabolism to fermentation, resulting in the conversion of methane to extracellular organic acids, which would serve as substrates for non-methanotrophic microbes. We intended to assess the prevalence of fermentation traits among freshwater gMOB. Therefore, we isolated two strains representing relevant freshwater gMOB genera, i.e., Methylovulum and Methylomonas, from boreal lakes, experimentally showed that they convert methane to organic acids and demonstrated via metagenomics that the fermentation potential is widely dispersed among lake and pond representatives of these genera. Combined with our recent study showing coherent results from another relevant freshwater gMOB genus, i.e., Methylobacter, we conclude that the conversion of methane to organic acids is a widely found trait among freshwater gMOB, highlighting their role as pivotal mediators of methane carbon into microbial food webs.

Organic matter lability modifies the vertical structure of methane-related microbial communities in lake sediments.

Eutrophication increases the input of labile, algae-derived, organic matter (OM) into lake sediments. This potentially increases methane (CH4) emissions from sediment to water through increased methane production rates and decreased methane oxidation efficiency in sediments. However, the effect of OM lability on the structure of methane oxidizing (methanotrophic) and methane producing (methanogenic) microbial communities in lake sediments is still understudied. We studied the vertical profiles of the sediment and porewater geochemistry and the microbial communities (16S rRNA gene amplicon sequencing) at five profundal stations of an oligo-mesotrophic, boreal lake (Lake Pääjärvi, Finland), varying in surface sediment OM sources (assessed via sediment C:N ratio). Porewater profiles of methane, dissolved inorganic carbon (DIC), acetate, iron, and sulfur suggested that sites with more autochthonous OM showed higher overall OM lability, which increased remineralization rates, leading to increased electron acceptor (EA) consumption and methane emissions from sediment to water. When OM lability increased, the abundance of anaerobic nitrite-reducing methanotrophs (Candidatus Methylomirabilis) relative to aerobic methanotrophs (Methylococcales) in the methane oxidation layer of sediment surface decreased, suggesting that Methylococcales were more competitive than Ca. Methylomirabilis under decreasing redox conditions and increasing methane availability due to their more diverse metabolism (fermentation and anaerobic respiration) and lower affinity for methane. Furthermore, when OM lability increased, the abundance of methanotrophic community in the sediment surface layer, especially Ca. Methylomirabilis, relative to the methanogenic community decreased. We conclude that increasing input of labile OM, subsequently affecting the redox zonation of sediments, significantly modifies the methane producing and consuming microbial community of lake sediments. IMPORTANCE Lakes are important natural emitters of the greenhouse gas methane (CH4). It has been shown that eutrophication, via increasing the input of labile organic matter (OM) into lake sediments and subsequently affecting the redox conditions, increases methane emissions from lake sediments through increased sediment methane production rates and decreased methane oxidation efficiency. However, the effect of organic matter lability on the structure of the methane-related microbial communities of lake sediments is not known. In this study, we show that, besides the activity, also the structure of lake sediment methane producing and consuming microbial community is significantly affected by changes in the sediment organic matter lability.

Characterization and genome analysis of a psychrophilic methanotroph representing a ubiquitous Methylobacter spp. cluster in boreal lake ecosystems.

Lakes and ponds are considered as a major natural source of CH4 emissions, particularly during the ice-free period in boreal ecosystems. Aerobic methane-oxidizing bacteria (MOB), which utilize CH4 using oxygen as an electron acceptor, are one of the dominant microorganisms in the CH4-rich water columns. Metagenome-assembled genomes (MAGs) have revealed the genetic potential of MOB from boreal aquatic ecosystems for various microaerobic/anaerobic metabolic functions. However, experimental proof of these functions, i.e., organic acid production via fermentation, by lake MOB is lacking. In addition, psychrophilic (i.e., cold-loving) MOB and their CH4-oxidizing process have rarely been investigated. In this study, we isolated, provided a taxonomic description, and analyzed the genome of Methylobacter sp. S3L5C, a psychrophilic MOB, from a boreal lake in Finland. Based on phylogenomic comparisons to MAGs, Methylobacter sp. S3L5C represented a ubiquitous cluster of Methylobacter spp. in boreal aquatic ecosystems. At optimal temperatures (3-12 °C) and pH (6.8-8.3), the specific growth rates (µ) and CH4 utilization rate were in the range of 0.018-0.022 h-1 and 0.66-1.52 mmol l-1 d-1, respectively. In batch cultivation, the isolate could produce organic acids, and the concentrations were elevated after replenishing CH4 and air into the headspace. Up to 4.1 mM acetate, 0.02 mM malate, and 0.07 mM propionate were observed at the end of the test under optimal operational conditions. The results herein highlight the key role of Methylobacter spp. in regulating CH4 emissions and their potential to provide CH4-derived organic carbon compounds to surrounding heterotrophic microorganisms in cold ecosystems.

Draft genome sequence data of a psychrophilic tundra soil methanotroph, Methylobacter psychrophilus Z-0021 (DSM 9914).

Psychrophilic methanotrophic bacteria are abundant and play an important role in methane removal in cold methanogenic environments, such as boreal and arctic terrestrial and aquatic ecosystems. They could be also applied in the bioconversion of biogas and natural gas into value-added products (e.g., chemicals and single-cell protein) in cold regions. Hence, isolation and genome sequencing of psychrophilic methanotrophic bacteria are needed to provide important data on their functional capabilities. However, psychrophilic methanotroph isolates and consequently their genome sequences are rare. Fortunately, Leibniz Institute, DSMZ-German Collection of Microorganisms and Cell Cultures GmbH was able to revive the long-extinct pure culture of a psychrophilic methanotrophic tundra soil isolate, Methylobacter psychrophilus Z-0021 (DSM 9914), from their stocks during 2022. Here, we describe the de novo assembled genome sequence of Methylobacter psychrophilus Z-0021 comprising a total of 4691082 bp in 156 contigs with a G+C content of 43.1% and 4074 coding sequences. The preliminary genome annotation analysis of Z-0021 identified genes encoding oxidation of methane, methanol and formaldehyde, assimilation of carbon and nitrate, and N2 fixation. In pairwise genome-to-genome comparisons with closely related methanotrophic strains, the strain Z-0021 had an average nucleotide identity (ANI) of 92.9% and 78.2% and a digital DNA-DNA hybridization (dDDH) value of 50.6% and 22% with a recently described psychrophilic, lake isolate, Methylobacter sp. S3L5C and a psychrotrophic, arctic wetland soil isolate, Methylobacter tundripaludum SV96, respectively. In addition, the respective similarities between genomes of the strains S3L5C and SV96 were 78.1% ANI and 21.8% dDDH. Comparison to widely used ANI and dDDH thresholds to delineate unique species (<95% ANI and <70% dDDH) suggests that Methylobacter psychrophilus Z-0021, Methylobacter tundripaludum SV96 and Methylobacter sp. S3L5C are different species. The draft genome of Z-0021 has been deposited at GenBank under the accession JAOEGU000000000.

Batch Experiments Demonstrating a Two-Stage Bacterial Process Coupling Methanotrophic and Heterotrophic Bacteria for 1-Alkene Production From Methane.

Methane (CH4) is a sustainable carbon feedstock for value-added chemical production in aerobic CH4-oxidizing bacteria (methanotrophs). Under substrate-limited (e.g., oxygen and nitrogen) conditions, CH4 oxidation results in the production of various short-chain organic acids and platform chemicals. These CH4-derived products could be broadened by utilizing them as feedstocks for heterotrophic bacteria. As a proof of concept, a two-stage system for CH4 abatement and 1-alkene production was developed in this study. Type I and Type II methanotrophs, Methylobacter tundripaludum SV96 and Methylocystis rosea SV97, respectively, were investigated in batch tests under different CH4 and air supplementation schemes. CH4 oxidation under either microaerobic or aerobic conditions induced the production of formate, acetate, succinate, and malate in M. tundripaludum SV96, accounting for 4.8-7.0% of consumed carbon from CH4 (C-CH4), while M. rosea SV97 produced the same compounds except for malate, and with lower efficiency than M. tundripaludum SV96, accounting for 0.7-1.8% of consumed C-CH4. For the first time, this study demonstrated the use of organic acid-rich spent media of methanotrophs cultivating engineered Acinetobacter baylyi ADP1 'tesA-undA cells for 1-alkene production. The highest yield of 1-undecene was obtained from the spent medium of M. tundripaludum SV96 at 68.9 ± 11.6 µmol mol Csubstrate -1. However, further large-scale studies on fermenters and their optimization are required to increase the production yields of organic acids in methanotrophs.

Characterization, genome analysis and genetic tractability studies of a new nanocellulose producing Komagataeibacter intermedius isolate.

Bacterial nanocellulose (BC) is a highly versatile biopolymer currently pursued as a material of choice in varied themes of biomedical and material science research fields. With the aim to extend the biotechnological applications, the genetic tractability of the BC producers within the Komagataeibacter genus and its potential as an alternative host chassis in synthetic biology have been extensively studied. However, such studies have been largely focused on the model Komagataeibacter spp. Here, we present a novel K. intermedius strain capable of utilizing glucose, and glycerol sources for biomass and BC synthesis. Genome assembly identified one bacterial cellulose synthetase (bcs) operon containing the complete gene set encoding the BC biogenesis machinery (bcsI) and three additional copies (bcsII-IV). Investigations on the genetic tractability confirmed plasmid transformation, propagation of vectors with pBBR1 and p15A origin of replications and constitutive and inducible induction of recombinant protein in K. intermedius ENS15. This study provides the first report on the genetic tractability of K. intermedius, serving as starting point towards future genetic engineering of this strain.

Simple enrichment system for hydrogen producers.

This study presents a simple enrichment system where gas pressure produced by microbes performs functions that are normally done by labor. The system was tested with Escherichia coli strains with different hydrogen production and growth capabilities. The results show that the system can enrich the best hydrogen producer.

Draft genome sequence data of methanotrophic Methylovulum psychrotolerans strain S1L and Methylomonas paludis strain S2AM isolated from hypoxic water column layers of boreal lakes.

Methanotrophic bacteria inhabit a wide range of natural (e.g. wetlands, lakes and soils) and anthropogenic (e.g. wastewater treatment plants and landfills) environments. They play a crucial role in mitigating atmospheric emissions of the greenhouse gas methane. There is also a growing interest in applying methanotrophs in the bioconversion of biogas - and natural gas - methane into value-added products (e.g. chemicals and single-cell protein). Hence, isolation and genome sequencing of methanotrophic bacteria is needed to provide important data on their functional capabilities. Here, we describe the de novo assembled draft genome sequences of Methylovulum psychrotolerans strain S1L isolated from hypoxic water column layer of boreal Lake Lovojärvi (Southern Finland), comprising total of 5090628 bp in 11 contigs with G+C - content of 50.9% and containing 4554 coding sequences. The draft genome of strain S1L represents the first published genome of M. psychrotolerans strain isolated from lake ecosystems. In addition, we present the genome sequence of Methylomonas paludis strain S2AM, isolated from water column of boreal Lake Alinen Mustajärvi (Southern Finland), comprising 3673651 bp in 1 contig with G+C - content of 48.2% and 3294 coding sequences. The draft genome of strain S2AM represents the first published genome of M. paludis. The preliminary genome annotation analysis of both S1L and S2AM identified genes encoding oxidation of methane, methanol, formaldehyde and formate, assimilation of carbon, ammonium and nitrate, N2 fixation, as well as various enzymes enabling the survival in hypoxic conditions, i.e. high-affinity oxidase, hemerythrins, fermentation enzymes (for production of acetate, succinate and H2) and respiratory nitrite reductases. The draft genomes have been deposited at GenBank under the accession JAGVVN000000000 for S1L and CP073754 for S2AM.

Characterization of Komagataeibacter Isolate Reveals New Prospects in Waste Stream Valorization for Bacterial Cellulose Production.

Komagataeibacter spp. has been used for the bioconversion of industrial wastes and lignocellulosic hydrolysates to bacterial cellulose (BC). Recently, studies have demonstrated the capacity of Komagataeibacter spp. in the biotransformation of inhibitors found in lignocellulosic hydrolysates, aromatic lignin-derived monomers (LDMs) and acetate. In general, detoxification and BC synthesis from lignocellulosic inhibitors requires a carbon flow from acetyl-coA towards tricarboxylic acid and gluconeogenesis, respectively. However, the related molecular aspects have not yet been identified in Komagataeibacter spp. In this study, we isolated a cellulose-producing bacterium capable of synthesizing BC in a minimal medium containing crude glycerol, a by-product from the biodiesel production process. The isolate, affiliated to Komagataeibacter genus, synthesized cellulose in a minimal medium containing glucose (3.3 ± 0.3 g/L), pure glycerol (2.2 ± 0.1 g/L) and crude glycerol (2.1 ± 0.1 g/L). Genome assembly and annotation identified four copies of bacterial cellulose synthase operon and genes for redirecting the carbon from the central metabolic pathway to gluconeogenesis. According to the genome annotations, a BC production route from acetyl-CoA, a central metabolic intermediate, was hypothesized and was validated using acetate. We identified that when K. rhaeticus ENS9b was grown in a minimal medium supplemented with acetate, BC production was not observed. However, in the presence of readily utilizable substrates, such as spent yeast hydrolysate, acetate supplementation improved BC synthesis.

Enhancing piezoelectric properties of bacterial cellulose films by incorporation of MnFe2O4 nanoparticles.

Low-cost and highly sensitive piezoelectric sensors were fabricated from bacterial cellulose (BC)/MnFe2O4 nanocomposite films via a co-precipitation method, followed by hot-pressing. MnFe2O4 nanoparticles were homogeneously distributed in the BC structure. The piezoelectric sensitivity measurements in the normal mode showed that the pristine BC film exhibited a sensitivity of ∼5 pC/N, whereas this value was increased to 23 pC/N for the composite film, which is comparable to the PVDF reference film. In the bending mode, the piezoelectric response increased to 25 pC/N and 57 pC/N for the BC film and the composite film, respectively. Moreover, the piezoelectric sensitivity was significantly enhanced using carbon tape electrodes attached directly to the films instead of sandwiched electrodes. This produced a sensitivity of greater than 50 pC/N for the MBC nanocomposite film in the normal mode measurement. Our work demonstrates the potential of using MBC composite films as inexpensive and highly sensitive flexible piezoelectric sensors.

Co-production of 1,3 propanediol and long-chain alkyl esters from crude glycerol.

Crude glycerol is an excellent carbon source for bacterial production systems. Bacterial fermentation often generates by-products that can offer an additional carbon pool to improve the product profile for optimal valorization. In this study, the properties of two phylogenetically distinct bacteria, Acinetobacter baylyi ADP1 and Clostridium butyricum, were coupled in a one-pot batch process to co-produce 1,3 propanediol (PDO) and long-chain alkyl esters (wax esters, WEs) from crude glycerol. In the process, A. baylyi deoxidized the growth medium allowing glycerol fermentation and PDO production by C. butyricum. Reaeration of the co-cultivations enabled A. baylyi to metabolize the fermentation by-products, acetate and butyrate, and synthesize intracellular WEs. To improve PDO production and A. baylyi growth, carbon and macronutrients in the growth medium were screened and optimized using Plackett-Burman and Box-Behnken models. The validation experiment revealed a good correlation between the observed and predicted values. The salting-out method recovered 89.5% PDO from the fermentation broth and in vacuo extraction resulted in a PDO content of 5.3 g L-1. Nuclear magnetic resonance revealed a WE content and yield of 34.4 ±â€¯1.4 mg L-1 and 34.2 ±â€¯3.2 mg WE g-1 dry cell weight, respectively. A molar yield of 0.65 mol PDO mol-1 and 0.62 µmol WE mol-1 crude glycerol was achieved with the synthetic consortium. This work emphasizes the strength of response surface methodology in improving production processes from the mutualistic association of divergent bacterial species in consortium. The co-production of PDO and WEs from crude glycerol is demonstrated for the first time in this study.

O2-requiring molecular reporters of gene expression for anaerobic microorganisms.

Many genetic reporter systems require molecular oxygen; therefore, the use of reporter genes to study molecular mechanisms in anaerobic microorganisms has been hampered by the lack of convenient reporting systems. We describe reporter gene whole cell-based biosensor systems based on luciferase genes and the associated oxygen-requiring enzymes. By using two different oxygen-dependent reporters, insect and bacterial luciferases, and two bacterial hosts, Gram (+) Bifidobacterium longum and Gram (-) Escherichia coli, we show that the enzymes can be used in gene expression studies of anaerobic bacteria. E. coli, a facultative anaerobe, was grown both in aerobic and anaerobic conditions with an arabinose-inducible expression system. We show that a short treatment time of few minutes in ambient atmosphere is sufficient to detect light emission from living cells that is directly proportional to the number of cells and to the inducer concentration. The induction levels were the same in both the aerobically and anaerobically cultured cells. Similar results were obtained in the case of B. longum cultured in anaerobic conditions.

Metabolic pairing of aerobic and anaerobic production in a one-pot batch cultivation.

BACKGROUND: The versatility of microbial metabolic pathways enables their utilization in vast number of applications. However, the electron and carbon recovery rates, essentially constrained by limitations of cell energetics, are often too low in terms of process feasibility. Cocultivation of divergent microbial species in a single process broadens the metabolic landscape, and thus, the possibilities for more complete carbon and energy utilization. RESULTS: In this study, we integrated the metabolisms of two bacteria, an obligate anaerobe Clostridium butyricum and an obligate aerobe Acinetobacter baylyi ADP1. In the process, a glucose-negative mutant of A. baylyi ADP1 first deoxidized the culture allowing C. butyricum to grow and produce hydrogen from glucose. In the next phase, ADP1 produced long chain alkyl esters (wax esters) utilizing the by-products of C. butyricum, namely acetate and butyrate. The coculture produced 24.5 ± 0.8 mmol/l hydrogen (1.7 ± 0.1 mol/mol glucose) and 28 mg/l wax esters (10.8 mg/g glucose). CONCLUSIONS: The cocultivation of strictly anaerobic and aerobic bacteria allowed the production of both hydrogen gas and long-chain alkyl esters in a simple one-pot batch process. The study demonstrates the potential of 'metabolic pairing' using designed microbial consortia for more optimal electron and carbon recovery.

Engineering and Characterization of Bacterial Nanocellulose Films as Low Cost and Flexible Sensor Material.

Some bacterial strains such as Komagataeibacter xylinus are able to produce cellulose as an extracellular matrix. In comparison to wood-based cellulose, bacterial cellulose (BC) holds interesting properties such as biodegradability, high purity, water-holding capacity, and superior mechanical and structural properties. Aiming toward improvement in BC production titer and tailored alterations to the BC film, we engineered K. xylinus to overexpress partial and complete bacterial cellulose synthase operon that encodes activities for BC production. The changes in cell growth, end metabolite, and BC production titers from the engineered strains were compared with the wild-type K. xylinus. Although there were no significant differences between the growth of wild-type and engineered strains, the engineered K. xylinus strains demonstrated faster BC production, generating 2-4-fold higher production titer (the highest observed titer was obtained with K. xylinus-bcsABCD strain producing 4.3 ± 0.46 g/L BC in 4 days). The mechanical and structural characteristics of cellulose produced from the wild-type and engineered K. xylinus strains were analyzed with a stylus profilometer, in-house built tensile strength measurement system, a scanning electron microscope, and an X-ray diffractometer. Results from the profilometer indicated that the engineered K. xylinus strains produced thicker BC films (wild type, 5.1 µm, and engineered K. xylinus strains, 6.2-10.2 µm). Scanning electron microscope revealed no principal differences in the structure of the different type BC films. The crystallinity index of all films was high (from 88.6 to 97.5%). All BC films showed significant piezoelectric response (5.0-20 pC/N), indicating BC as a promising sensor material.

Recombinant antibodies for specific detection of clostridial [Fe-Fe] hydrogenases.

Biological hydrogen production is based on activity of specific enzymes called hydrogenases. Hydrogenases are oxygen sensitive metalloenzymes containing Ni and/or Fe atoms at the active site, catalyzing reversible reduction of protons. Generally, [Fe-Fe] hydrogenases prefer proton reduction to molecular hydrogen, a potential energy carrier molecule that can be produced by bioprocesses in sustainable manner. Thus, monitoring tools have been developed to study the relationship between [Fe-Fe] hydrogenases and biohydrogen production in bioreactors at DNA and RNA levels. In the present study, novel molecular tools are introduced for quantitative monitoring of clostridial [Fe-Fe] hydrogenases at the protein level. Aerobic and anaerobic biopanning (for inactive and active [Fe-Fe] hydrogenase, respectively) of phage displayed single-chain variable fragment (scFv) antibody libraries aided in isolating nine potential scFvs. The enriched antibodies demonstrated high specificity towards Clostridium spp. [Fe-Fe] hydrogenases allowing detection from pure and mixed cultures. Additionally, the antibodies showed different binding characteristics towards hydrogenase catalytic states, providing a possible means for functional detection of clostridial [Fe-Fe] hydrogenases. From hydrogenase-antibody interaction studies we observed that though antibody binding reduced the enzyme catalytic activity, it facilitated to retain hydrogen evolution from oxygen exposed hydrogenases.

Metabolic engineering of Acinetobacter baylyi ADP1 for removal of Clostridium butyricum growth inhibitors produced from lignocellulosic hydrolysates.

BACKGROUND: Pretreatment of lignocellulosic biomass can produce inhibitory compounds that are harmful for microorganisms used in the production of biofuels and other chemicals from lignocellulosic sugars. Selective inhibitor removal can be achieved with biodetoxification where microorganisms catabolize the inhibitors without consuming the sugars. We engineered the strictly aerobic Acinetobacter baylyi ADP1 for detoxification of lignocellulosic hydrolysates by removing the gene for glucose dehydrogenase, gcd, which catalyzes the first step in its glucose catabolism. RESULTS: The engineered A. baylyi ADP1 strain was shown to be incapable of consuming the main sugar components of lignocellulosic hydrolysates, i.e., glucose, xylose, and arabinose, but rapidly utilized acetate and formate. Formate was consumed during growth on acetate and by stationary phase cells, and this was enhanced in the presence of a common aromatic inhibitor of lignocellulosic hydrolysates, 4-hydroxybenzoate. The engineered strain tolerated glucose well up to 70 g/l, and the consumption of glucose, xylose, or arabinose was not observed in prolonged cultivations. The engineered strain was applied in removal of oxygen, a gaseous inhibitor of anaerobic fermentations. Co-cultivation with the A. baylyi ADP1 gcd knockout strain under initially aerobic conditions allowed the strictly anaerobic Clostridium butyricum to grow and produce hydrogen (H2) from sugars of the enzymatic rice straw hydrolysate. CONCLUSIONS: We demonstrated that the model organism of bacterial genetics and metabolism, A. baylyi ADP1, could be engineered to be an efficient biodetoxification strain of lignocellulosic hydrolysates. Only one gene knockout was required to completely eliminate sugar consumption and the strain could be used in production of anaerobic conditions for the strictly anaerobic hydrogen producer, C. butyricum. Because of these encouraging results, we believe that A. baylyi ADP1 is a promising candidate for the detoxification of lignocellulosic hydrolysates for bioprocesses.

Bioprocess data mining using regularized regression and random forests.

BACKGROUND: In bioprocess development, the needs of data analysis include (1) getting overview to existing data sets, (2) identifying primary control parameters, (3) determining a useful control direction, and (4) planning future experiments. In particular, the integration of multiple data sets causes that these needs cannot be properly addressed by regression models that assume linear input-output relationship or unimodality of the response function. Regularized regression and random forests, on the other hand, have several properties that may appear important in this context. They are capable, e.g., in handling small number of samples with respect to the number of variables, feature selection, and the visualization of response surfaces in order to present the prediction results in an illustrative way. RESULTS: In this work, the applicability of regularized regression (Lasso) and random forests (RF) in bioprocess data mining was examined, and their performance was benchmarked against multiple linear regression. As an example, we used data from a culture media optimization study for microbial hydrogen production. All the three methods were capable in providing a significant model when the five variables of the culture media optimization were linearly included in modeling. However, multiple linear regression failed when also the multiplications and squares of the variables were included in modeling. In this case, the modeling was still successful with Lasso (correlation between the observed and predicted yield was 0.69) and RF (0.91). CONCLUSION: We found that both regularized regression and random forests were able to produce feasible models, and the latter was efficient in capturing the non-linearity in the data. In this kind of a data mining task of bioprocess data, both methods outperform multiple linear regression.