Predicted Relative Metabolomic Turnover (PRMT): determining metabolic turnover from a coastal marine metagenomic dataset
© Larsen et al; licensee BioMed Central Ltd. 2011
Received: 3 January 2011
Accepted: 14 June 2011
Published: 14 June 2011
The world's oceans are home to a diverse array of microbial life whose metabolic activity helps to drive the earth's biogeochemical cycles. Metagenomic analysis has revolutionized our access to these communities, providing a system-scale perspective of microbial community interactions. However, while metagenome sequencing can provide useful estimates of the relative change in abundance of specific genes and taxa between environments or over time, this does not investigate the relative changes in the production or consumption of different metabolites.
We propose a methodology, Predicted Relative Metabolic Turnover (PRMT) that defines and enables exploration of metabolite-space inferred from the metagenome. Our analysis of metagenomic data from a time-series study in the Western English Channel demonstrated considerable correlations between predicted relative metabolic turnover and seasonal changes in abundance of measured environmental parameters as well as with observed seasonal changes in bacterial population structure.
The PRMT method was successfully applied to metagenomic data to explore the Western English Channel microbial metabalome to generate specific, biologically testable hypotheses. Generated hypotheses linked organic phosphate utilization to Gammaproteobactaria, Plantcomycetes, and Betaproteobacteria, chitin degradation to Actinomycetes, and potential small molecule biosynthesis pathways for Lentisphaerae, Chlamydiae, and Crenarchaeota. The PRMT method can be applied as a general tool for the analysis of additional metagenomic or transcriptomic datasets.
Marine biomes dominate the planet's surface and single-celled microorganisms are responsible for up to 98% of the ocean's primary productivity ; understanding the nutrient and carbon cycles of the world's oceans has key applications for understanding global ecology. The extremely diverse marine microbial communities mediate the largest active pool of near-surface carbon on the planet  and are a dominant force in the planet's biogeochemical cycles . The L4 Station of the Western Channel Observatory (WCO), an oceanographic time-series and marine biodiversity reference site in the Western English Channel http://www.westernchannelobservatory.org.uk, provides a unique opportunity to study a coastal marine microbial ecosystem. Environmental parameter data from the WCO have been continuously monitored for over a century. More recently, microbial metagenomic data collected from this site have shown that the abundance and relative composition of genes and taxa change over time, demonstrating seasonal structure and predictable community responses to environmental parameters [4–7].
The seasonal structure in the community composition of both taxa and genes has potential repercussions for the seasonal succession in metabolic potential, which will drive the range and relative abundance of metabolites produced and consumed by a community. Metagenomic analyses explore the functional potential of an ecosystem by describing the changes in the abundance of genes annotated with unique enzyme functions. Here we propose a methodology that alters this paradigm, by describing the metagenome in terms of the relative change in the production or consumption of specific metabolites. Instead of exploring gene-space using an environmental metagenomic analysis, this study explores derived metabolite-space, inferred from the metagenome. The predicted environmental metabolome is the set of all detected unique enzyme functions encoded in a metagenome and all of the metabolites implied by those activities. This network of predicted metabolic reactions represents the theoretical metabolic potential of an environmental metabolome and provides a novel means through which metagenomic observations of seasonal and/or biogeographic trends in microbial communities can be utilized to explore community-wide metabolic dynamics.
Understanding the metabolomic interactions in any marine microbial community is a daunting task. Classical metabolomic techniques such as NMR or GC-MS [8–11], while powerful, provide measurements for just a fraction of the metabolites predicted to be present in a metabolome. As it is not possible to easily measure every metabolite in an environment, it is important to determine the most pertinent parameters, those that will allow investigators to generate testable biological hypotheses. Currently, methods are available to extract a wide variety of environmental features from metagenomic sequence data [12–14]. Metagenomics can be used to determine the taxonomic and functional diversity of a microbial community via automated pipelines  to curated protein databases such as RefSeq , KEGG , KEGGnoggs , SEED , PFAM  or TIGRfam , and linking environmental conditions with specific metabolic activities inferred from metagenomic data .
Here, we propose a novel computational method, Predicted Relative Metabolic Turnover (PRMT), which enables comparative analyses of environmental metabolomes predicted from metagenomic data. To support the methodology with specific biological observations, PRMT was applied to a metagenomic dataset from the time-series study of the bacterial environmental metabolome in the Western English Channel . The PRMT approach also correctly predicted considerable correlations between environmental metabolite concentrations and PRMT metabolomic predictions. It also correctly predicting seasonal variations in the bacterial community primary productivity, and generated specific, testable biological hypotheses for organic phosphate utilization by Gammaproteobactaria, Plantcomycetes, and Betaproteobacteria; chitin degradation by Actinomycetes; and potential small molecule biosynthesis pathways for Lentisphaerae, Chlamydiae, and Crenarchaeota.
Materials and methods
Data used for validation of methodology
Metagenomic sequencing data was acquired from a recent study examining the seasonal structure of functional potential in the Western English Channel . Specific data points for use in this study came from day/night pairs of samples taken on January 28, April 27, and August 27 2008. All data collected on the same day were averaged; previous analyses of these data have shown that the metagenomic functional gene profile is statistically identical between day and night in each of these samples . All metagenomic data were annotated with MG-RAST  using parameters previously described . Specifically, nucleic acid sequences were excluded if annotated as rRNA, and all subsequent reads were annotated against the SEED database using MG-RAST (e-value < 1 × 10-3; minimum length of alignment of 50 bp; minimum sequence nucleotide identity of 50%; ) to produce an abundance matrix of functional genes and protein-derived predicted taxonomies across the DNA samples. Short read data are available through the European Nucleotide Archive (ENA) short read archive under ERP000118 http://www.ebi.ac.uk/ena/data/view/ERP000118. All data are available on the CAMERA website under 'Western Channel Observatory Microbial Metagenomic Study' http://camera.calit2.net. MG-RAST annotations of metagenomic data were collected under IDs 4445064.3, 444077.3, 4445065.3, 4445066.3, 4445068.3, 4444083.3, 4445069.3, and 4445070 http://metagenomics.anl.gov/. All submissions conform to the minimum information standards (MIxS) of the Genomic Standards Consortium .
PRMT analysis approach
PRMT scores predict the change in turnover of metabolites (defined as the potential for consumption or production) in an environmental metabolome, given the relative abundance of genes for unique enzyme functions detected in different metagenomes. In this manuscript, we use the term "unique enzyme function" to describe a specific annotation applied to an enzyme, i.e. "Phosphotransferases with an alcohol group as acceptor". We use "enzyme reactions" to refer to metabolite transformations catalyzed by an enzyme function, i.e "ATP + D-Glycerate ↔ ADP + 3-Phospho-D-glycerate". A unique enzyme function may catalyze more than one enzyme reaction and an enzyme reaction may be catalyzed by more than one unique enzyme function. A metabolite is a molecular compound that is a reactant or product in an enzyme reaction. In PRMT, a metabolite is never the protein product of a gene in the metagenome.
(1) Generate Environmental Metabolome Matrix (EMM). A network, constructed from annotated unique enzyme functions, enzyme reactions and inferred metabolites, is used to generate the predicted environmental metabolome; this network is expressed as a connectivity matrix, the Environmental Metabolomic Matrix (EMM; Figure 1). The resulting EMM matrix has dimensions n x m, where n is the number of predicted metabolites in the environmental metabolome and m is the number of unique enzyme functions detected in the set of analyzed metagenomes. The set of Kyoto Encyclopedia of Genes and Genomes (KEGG)  metabolic reactions was used to represent the set of possible enzyme reactions in an environmental metabolome and Enzyme Commission (EC) number annotations for enzyme activity from MG-RAST annotations of metagenomic sequences were used to assign unique enzyme functions to the predicted proteins encoded in a metagenome. KEGG reactions used were selected exclusively from KEGG 'Metabolism' pathways. KEGG pathway maps that are not directly related to specific metabolic activities in microorganisms (e.g. Human Diseases, Drug Development, etc.) were not used in generation of metabolomes. A complete list of the 156 KEGG pathways used in construction of EMM is provided as supplementary data [Additional file 1, Table S1]. As is common in network analysis of biochemical systems , the KEGG metabolites water, di- and tri-phosphonucleotides, and all ubiquitous cofactors were excluded from the list of possible reactants as non-specific to particular reactions and metabolic processes. If a reaction is identified as reversible in the KEGG database, then both forward and reverse reactions are included in the EMM. If a particular metabolite transformation is attributed to more than one enzyme activity, then each transformation reaction is incorporated into the EMM. The stoichiometry of each reaction was not considered as the quantity of metabolites was not considered, just the relative turnover of each.
EAC i is the Enzyme Activity Count for enzymatic function i, and N seq,i is the number of sequence reads in the metagenome annotated with a unique enzyme function i. Collected EACs for a set of metagenomes are normalized (nEAC) by quantiles . Quantile normalization is a technique for making distributions from multiple datasets identical in statistical properties. To quantile-normalize the sets of EACs, the EACs are sorted largest to smallest. A reference distribution is made from the sorted lists such that the highest value in all cases becomes the mean of the highest values, the second highest value becomes the mean of the second highest values, and so on. For each set of EACs, nEAC is generated by assigning the distributions of the reference distribution to the observed EAC distribution. The set of nEACs for a metagenome is expressed as a vector of length n, where n is the total number of unique enzyme functions found in the set of metagenomes.
is a vector of PRMT-scores of length m, where m is the number of metabolic compounds in the EMM. and are vectors of normalized enzyme gene counts (nEAC) of length n, where n is the number of unique enzyme functions annotated to metagenomic sequences, and x and y refer to different metagenome datasets. M is the EMM, a matrix of dimensions m x n. Changes in nEAC for a reversible interaction do not change calculated PRMT-scores for metabolites. For analysis of the L4 WCO environmental metabolome, the reference nEAC was the average of EAC's across all samples from January, April, and August.
Interpretation of PRMT scores
PRMT-scores are unit-less values that represent the change of the turnover of a metabolite in a predicted metabolome relative to a reference metabolome. A PRMT-score is calculated for every metabolite in the EMM for a metagenome. The value and sign (positive or negative) of a PRMT-score provides information about a metabolite's relative turnover. Although a thorough interpretation of a PRMT-score requires that it be considered in the greater context of the complete network, it can be broadly interpreted as follows: A positive PRMT score predicts increased metabolic turnover and relatively greater consumption of a metabolite. A negative PRMT-score predicts decreased turnover and relatively greater accumulation of a metabolite. It is important to note that PRMT-scores do not predict net production or consumption of a metabolite.
Correlations between PRMT-scores and environmental parameters
Correlations of calculated PRMT scores with relative abundance of selected environmental parameter measurements.
Total Organic Nitrogen
Total Organic Carbon
Soluble Reactive Phosphorus
x is a measured experimental parameter and is the average for the parameter across all samples.
PCCs were calculated between four different combinations, namely; all measured environmental metabolites and metagenomic reads annotated to SEED subsystems; environmental metabolites and bacterial phyla; environmental metabolites and PRMT-scores; and PRMT-scores and bacterial phyla. It was considered that statistical significance could not be reliably assigned due to the small number of samples used in the analysis (3 seasonal time points). This was of particular concern given the method's reliance on a multiple-test based procedure. In order to address these concerns, and to provide an informal confidence estimate with which to judge each individual PCC, 10,000 randomized re-samplings of the initial data were used to generate a distribution of PCC-scores. Observed PCC-scores that were in the top or bottom 5th percentile of randomized re-sampled were considered to be a strong correlation.
A graphical representation of observed correlations was performed using 'Cytoscape'  to generate a network in which experimental measurements are represented as nodes and strong correlations between relative abundance of measurements were represented as edges.
Results and Discussion
Correlation networks between environmental parameters, bacterial phyla, and SEED subsystems
A strong negative correlation was observed between the relative abundance of the SEED subsystem 'Photosynthesis metabolism' and the concentration of chlorophyll A. This is similar to previous observations  that indicate the relative abundance of genes with homology to components of the cyanobacterial photosynthesis pathway peak in winter, while chlorophyll A concentrations, which relate to Eukaryotic phytoplankton abundance, peak in summer. There is also a strong negative correlation between the relative change in concentration of SRP and the SEED subsystem for 'Phosphate metabolism'. This suggests that when there are more genes for phosphate metabolism there is a greater predicted likelihood for SRP consumption. As a caveat for all of these relationships, correlation does not prove causality; they do however, identify relationships that invite further investigation.
The analysis of the relationships between phyla, environmental parameters, and functional genes indicates that changes in taxonomic diversity affect environmental metabolomic functional potential. To complement this gene-centric perspective, we present a metabolite-centric tool, PRMT, which infers metabolic activity from unique enzyme functions.
The Environmental Metabolomic Matrix from the Western English Channel
PRMT scores calculated from the Western English Channel EMM
PRMT scores were calculated for each predicted metabolite in the EMM at each of three seasonal time points (January 28, April 27, and August 27, 2008). Correlations between PRMT-scores and relative abundance of environmental parameters were generated (Table 1). The complete set of PRMT-scores is available as supplemental data (Additional file 3, Table S2). A strong negative correlation between PRMT scores and relative changes in the concentration of an environmental parameter demonstrates that the metabolomic capacity for synthesis of a parameter increased when the parameter is at a relatively higher concentration. Conversely, a strong positive correlation occurs if metabolomic capacity for the consumption of a metabolite increases when the environmental parameter is at a relatively higher concentration. Only three of the six correlations were considered strong (i.e. in the top or bottom 5th percentile of randomized resamples). However, using a calculation for the cumulative normal distribution (CND), given the distribution of correlations between relative abundances of environmental parameters and all calculated PRMT-scores, the probability that all six PCC-scores would be between -1 and -0.8 was < 1 × 10-256 (lowest is -0.81; Table 1). This probability indicates that, even though some individual correlations are not strong (based on the given criteria), it is extremely unlikely that all six correlations could all be between -1 and -0.8 by chance, assuming a normal distribution of correlation coefficients.
Correlations between PRMT-scores and the relative abundances of bacterial phyla can be used to generate hypotheses regarding the potential of different taxonomic groups to produce specific metabolites in subnetworks in the EMM. For example, Actinobacteria play an important environmental role in the decomposition of cellulose . The relative abundance of Actinobacteria had a strong negative correlation with PRMT-scores for the predicted metabolite cellulose. This strong negative correlation indicates that when the Actinobacteria are relatively more abundant, the predicted tendency for cellulose to be consumed in the environmental metabolome increases. In another example, the Planctomycetes was the only phyla whose change in relative abundance demonstrated a strong positive correlation with PRMT-scores for L-Glutamate. This was expected, as the cell wall scaffolds of most Plancotomycetes are made up of glycoproteins rich in glutamate instead of murein, as is common for other bacterial taxa . Of greater interest than investigating the predicted metabolome for individual metabolites is to seek connected metabolic subnetworks in the EMM that have strong correlations with changes in relative abundance of specific taxa. Connected subnetworks imply not just single metabolites, but partial or complete biochemical pathways. These connected subnetworks were generated by identifying all edges in the network that connect pairs of metabolite nodes where both nodes' PRMT-scores significantly correlate with the relative abundance of a specific bacterial phylum (Figure 3). The largest connected subnetwork for interactions that correlated with specific phyla was comprised of 28 metabolites and 70 enzyme reactions, and was comprised of metabolic interactions associated with the metabolism of phosphonoacetate. PRMT-scores for metabolites in this largest subnetwork strongly correlated with relative abundances of Planctomycetes and Gammaproteobacteria. In the set of connected subnetworks were 5 small subnetworks of 5 or 6 nodes, each with PRMT-scores that exclusively correlated to the relative abundance of Lentisphaerae. Predicted metabolites in these subnetworks were associated with flavanoids and anthocyanins. Flavonoids have been observed to have antimicrobial effects on marine microorganisms  and flavonoid biosynthesis can be hypothesized to either have defense-related functions in Lentispharae or else other marine microorganisms possess the capacity to synthesize flavonoids in response to increased relative abundances of Lentispharae. Biosynthesis of terpanoids from the precursor Farnesyl pyrophosphate correlated strongly, not only with the relative abundances of Crenarchaeota and Chlamydiae, but also with the relative change in concentration of TOC. Terpenoids are a functionally diverse set of molecules whose synthesis by marine bacteria has been previously reported [29, 30].
Specific Examples of PRMT Analysis: primary production, phosphonate metabolism, and chitin catabolism
The PRMT scores for CO2 were most positive in August (indicating consumption) and most negative in January (indicating synthesis) (Figure 4). Also, PRMT scores indicated increased production of alpha amino acids (biomass) in summer, and active consumption in the winter (Figure 4). This matches expectations, as both bacterial and eukaryotic cellular biomasses are higher in summer . When compared to investigating the abundance of sequences with homology to photosynthetic genes, which predict that the photosynthetic apparatus is more abundant in winter , PRMT scores reveal another level of potentially more useful information about the productivity of the ecosystem. While generated from the same data, consideration of sequence abundances and PRMT scores provide two unique perspectives. The largest PRMT-score shift in primary production was observed for nitrate. PRMT-scores for nitrite correlated strongly with the observed relative abundance of chlorophyll A. PRMT-scores for orthophosphate and alpha-amino acid all had strong correlations with PRMT-scores for CO2. As previously noted (Table 1), relative abundance of biological measurements of chlorophyll A had a strong correlation with PRMT-scores for chlorophyll A, linking the molecular components for photosynthesis with the PRMT-predicted capacity to synthesize the mechanism of photosynthesis.
One key limiting metabolite for primary productivity-derived biomass is iron [33, 34]. Increased concentrations of Fe2+ correlates with increased primary productivity in ocean microbial populations [35–37]. The PRMT-scores for Fe2+ demonstrated a strong positive correlation with PRMT-scores for alpha amino-acids, the EMM proxy for biomass, as well as the relative abundance of measured TON. PRMT analysis identified a sound positive correlation between Fe2+ and production of bacterial biomass.
Phosphonoacetate metabolism: Soluble Reactive Phosphorus (SRP) is considered to be a limited nutrient in marine ecosystems, although this is probably not the case in the Western English Channel . However, it was interesting that PRMT scores for inorganic phosphate (orthophosphate) were strongly negatively correlated (PCC -0.93) with changes in the relative concentration SRP (Table 1). One reason for this discrepancy is that SRP is comprised of both inorganic and organic phosphate compounds. One group of organic phosphate compounds, the phosphonates, were considered recalcitrant to biological life for many years, but recent evidence from the Western English Channel suggests that they are used by a wide variety of bacteria and constitute a large fraction of the available phosphorus pool . Currently, it is not technically feasible to measure specific phosphonate concentrations in marine systems, and this makes computational approaches like PRMT potentially valuable for investigating their relative turnover. Strikingly, 92% of the enzyme activities in the phosphonate and phosphinate metabolism KEGG pathway (map00440) were represented in the predicted EMM, indicating that the metabolic capacity to utilize phosphonates was present in the metagenomes.
The PRMT-scores for phosphonate compound, 3-Phosphonopyruvate, had strong correlations with PRMT-scores for CO2, and with biological measurements of chlorophyll A, suggesting a potential relationship between phosphonopyruvate metabolism and primary production. This is not totally unexpected, as phosphonopyruvate is a key intermediate step in the biosynthesis of all known natural phosphonates . Another phosphonate of potential interest is phosphonoacetate, as the presence of phosphonoacetate as a natural product has been suspected but not confirmed . Measured phosphonoacetate was strongly correlated with relative abundances of Planctomycetes, Gamaproteobacteria, and Betaproteobacteria. This insight into metabolic pathways associated with a metabolite difficult to measure in the environment led to a study to screen for a phosphonoacetate-oxidising activity in marine microorganisms; such an activity has recently been identified in cell-extracts of Roseovarius nubinhibens, and is now being characterized [unpublished data].
These reactions were all represented in the Western English Channel metagenomes. In the KEGG metabolic pathways, Fructose-6P metabolism leads to KEGG pathways for Glycolysis/gluconeogenesis (map00010) and Fructose and mannose metabolism (map00051), where both acetate and ammonia are generated.
We have presented a novel computational approach, Predictive Relative Metabolic Turnover (PRMT), for making predictions of the relative change in the production or consumption of a metabolite by a microbial community. The set of all PRMT-scores for all the metabolites predicted to be present in an environmental metabolome provides a system-scale representation of the environmental metabolome derived from metagenomic sequence analysis that changes in response to fluctuations in the structure of its bacterial community. The approach was supported by comparing predicted relative turnover of metabolites to relative abundance of the biological measurements for those same metabolites. PRMT also provided the expected relationships between CO2, Iron, orthophosphate, nitrate, and chlorophyll with marine primary production. Specific, testable, biological hypotheses regarding the utilization of organic phosphorus and chitin were made, some of which are currently being actively investigated. The predictions made by PRMT are consistent with the hypothesis that bacterial population diversity is linked to the metabolic capacity of the community. While we have restricted our analysis to the metagenome, PRMT calculations are equally applicable to metatranscriptomic data. Ongoing environmental monitoring projects such as the Global Ocean Survey, TARA Oceans, Hawaiian Ocean Time Series, Bermudan Ocean Time Series, the Long Term Ecological Research sites, NEON, and the extended application of the Earth Microbiome Project http://www.earthmicrobiome.org are generating vast amounts of metagenomic and experimental metadata that can readily be investigated by further PRMT analyses.
Predicted Relative Metabolomic Turnover
Environmental Metabolomic Matrix
Enzyme Activity Count
Normalized Enzyme Activity Count
Total Organic Carbon
Total Organic Nitrogen
Pearson Correlation Coefficient.
The work was funded by the US Department of Energy, Office of Science, Office of Biological and Environmental Research, Biological Systems Science Division under contract DE-AC02-06CH11357 at Argonne, IL. Sequence data was produced with a grant from the Natural Environmental Research Council (NERC - NE/F00138X/1).
- Jørgensen BB, Boetius A: Feast and famine--microbial life in the deep-sea bed. Nat Rev Microbiol. 10: 770-81.
- Buesseler KO, Lamborg CH, Boyd PW, Lam PJ, Trull TW, Bidigare RR, Bishop JK, Casciotti KL, Dehairs F, Elskens M, Honda M, Karl DM, Siegel DA, Silver MW, Steinberg DK, Valdes J, Van Mooy B, Wilson S: Revisiting carbon flux through the ocean's twilight zone. Science. 2008, 316 (5824): 567-70.View ArticleGoogle Scholar
- Paul Falkowski, Tom Fenchel, Edward Delong: The Microbial Engines That Drive Earth's Biogeochemical Cycles. Science. 2008, 320 (5879): 1034-1039. 10.1126/science.1153213.View ArticleGoogle Scholar
- Gilbert JA, Field D, Swift P, Thomas S, Cummings D, Temperton B, Weynberg K, Huse S, Hughes M, Joint I, Somerfield PJ, Mühling M: The Taxonomic and Functional Diversity of Microbes at a Temperate Coastal Site: A 'Multi-Omic' Study of Seasonal and Diel Temporal Variation. PLoS ONE. 2010, 5 (11): e15545-10.1371/journal.pone.0015545.PubMed CentralView ArticlePubMedGoogle Scholar
- Gilbert JA, Meyer F, Joint I, Mühling M, Field D: Short Metagenome Report: Metagenomes and metatranscriptomes from the L4 long term coastal monitoring station in the Western English Channel. SIGS. 2010, 3 (2):
- Mitra S, Gilbert JA, Field D, Huson DH: Comparison of multiple metagenomes using phylogenetic networks based on ecological indices. ISME J. 2010, 4: 1236-1242. 10.1038/ismej.2010.51.View ArticlePubMedGoogle Scholar
- Gilbert JA, Field D, Swift P, Newbold L, Oliver A, Smyth T, Somerfield P, Huse S, Joint I: Seasonal succession of microbial communities in the Western English Channel using 16S rDNA-tag pyrosequencing. Env Microb. 2009, 11 (12): 3132-3139. 10.1111/j.1462-2920.2009.02017.x.View ArticleGoogle Scholar
- Bundy JG, Davey MP, Viant MR: Environmental metabolomics: A critical review and future perspectives. Metabolomics. 2009, 5: 3-21. 10.1007/s11306-008-0152-0.View ArticleGoogle Scholar
- Viant MR: Metabolomics of aquatic organisms: the new omics on the block. Mar. Ecol. Prog. Series. 2007, 332: 301-306.View ArticleGoogle Scholar
- Viant MR: Recent developments in environmental metabolomics. Molecular Biosystems. 2008, 4: 980-986. 10.1039/b805354e.View ArticlePubMedGoogle Scholar
- Lin C-Y, Viant MR, Tjeerdema RS: Metabolomics: Methodologies and Applications in the Environmental Sciences. J. Pestic. Sci. 2006, 31: 245-251. 10.1584/jpestics.31.245.View ArticleGoogle Scholar
- Wooley JC, Godzik A, Friedberg I: A Primer on Metagenomics. PLoS Comput Biol. 2010, 6 (2): e1000667-10.1371/journal.pcbi.1000667.PubMed CentralView ArticlePubMedGoogle Scholar
- Heidelberg KB, Gilbert JA, Joint I: Review: The revolution in genomic approaches to describe environmental microbial diversity. Microb Biotechnol. 2010, 3 (5): 531-543. 10.1111/j.1751-7915.2010.00193.x.PubMed CentralView ArticlePubMedGoogle Scholar
- Gilbert JA, Dupont C: Marine Metagenomics. Annual Review of Marine Science. 2011, 3: 347-371. 10.1146/annurev-marine-120709-142811.View ArticlePubMedGoogle Scholar
- Meyer F, Paarmann D, D'Souza M, Olson R, Glass EM, Kubal M, Paczian T, Rodriguez A, Stevens R, Wilke A, Wilkening J, Edwards RA: The metagenomics RAST server - a public resource for the automatic phylogenetic and functional analysis of metagenomes. BMC Bioinformatics. 2008, 9: 386-10.1186/1471-2105-9-386.PubMed CentralView ArticlePubMedGoogle Scholar
- Pruitt KD, Maglott DR: NCBI Reference Sequence (RefSeq): a curated non-redundant sequence database of genomes, transcripts and proteins. Nucleic Acids Res. 2007, D61-5. 35 Database
- Kanehisa M, Goto S: KEGG: Kyoto Encyclopedia of Genes and Genomes. Nucleic Acids Res. 2000, 28: 27-30. 10.1093/nar/28.1.27.PubMed CentralView ArticlePubMedGoogle Scholar
- Muller J, Szklarczyk D, Julien P, Letunic I, Roth A, Kuhn M, Powell S, von Mering C, Doerks T, Jensen LJ, Bork P: KeggNOG v2.0: extending the evolutionary genealogy of genes with enhanced non-supervised orthologous groups, species and functional annotations. Nucleic Acids Res. 2010, D190-5. Database
- Overbeek R, Begley T, Butler RM, Jomuna O, Choudhuri V, Chuang H, Cohoon M, de Crécy-Lagard V, Diaz N, Disz T, Edwards R, Fonstein M, Frank ED, Gerdes S, Glass EM, Goesmann A, Hanson A, Iwata-Reuyl D, Jensen R, Jamshidi N, Krause L, Kubal M, Larsen N, Linke B, McHardy AC, Meyer F, Neuweger H, Olsen G, Olson R, Osterman A, Portnoy V, Pusch GD, Rodionov DA, Rückert C, Steiner J, Stevens R, Thiele I, Vassieva O, Ye Y, Zagnitko O, Vonstein V: The subsystems approach to genome annotation and its use in the project to annotate 1000 genomes. Nucleic Acids Res. 2005, 17: 5691-702.View ArticleGoogle Scholar
- Finn RD, Tate J, Mistry J, Coggill PC, Sammut SJ, Hotz HR, Ceric G, Forslund K, Eddy SR, Sonnhammer EL, Bateman A: The Pfam protein families database. Nucleic Acids Res. 2008, D281-8. Database
- Selengut JD, Haft DH, Davidsen T, Ganapathy A, Gwinn-Giglio M, Nelson WC, Richter AR, White O: TIGRFAMs and Genome Properties: tools for the assignment of molecular function and biological process in prokaryotic genomes. Nucleic Acids Res. 2007, 35 (Database): D260-4. 10.1093/nar/gkl1043.PubMed CentralView ArticlePubMedGoogle Scholar
- Gianoulis TA, Raes J, Patel PV, Bjornson R, Korbel JO, Letunic I, Yamada T, Paccanaro A, Jensen LJ, Snyder M, Bork P, Gerstein MB: Quantifying environmental adaptation of metabolic pathways in metagenomics. Proc Natl Acad Sci USA. 2009, 106 (5): 1374-9. 10.1073/pnas.0808022106. Epub 2009 Jan 22PubMed CentralView ArticlePubMedGoogle Scholar
- Field D, Joint I, Gilbert JA, et al: Towards a richer description of our complete collection of genomes and metagenomes: the "Minimum Information about a Genome Sequence" (MIGS) specification. Nat. Biotech. 2008, 6 (5): 541-7.View ArticleGoogle Scholar
- Ma HW, Aeng AP: Reconstruction of metabolic networks from genome data and analysis of their global structure for various organisms. Bioinformatics. 2003, 19 (2): 270-277. 10.1093/bioinformatics/19.2.270.View ArticlePubMedGoogle Scholar
- Rao Y, Lee Y, Jarjoura D, Ruppert AS, Liu CG, Hsu JC, Hagan JP: A comparison of normalization techniques for microRNA microarray data. Stat Appl Genet Mol Biol. 2008, 7 (1): Article22. Epub 2008 Jul 21Google Scholar
- Petković H, Cullum J, Hranueli D, Hunter IS, Perić-Concha N, Pigac J, Thamchaipenet A, Vujaklija D, Long PF: Genetics of Streptomyces rimosus, the oxytetracycline producer. Microbiol Mol Biol Rev. 2006, 70 (3): 704-28. 10.1128/MMBR.00004-06.PubMed CentralView ArticlePubMedGoogle Scholar
- Glöckner FO, Kube M, Bauer M, Teeling H, Lombardot T, Ludwig W, Gade D, Beck A, Borzym K, Heitmann K, Rabus R, Schlesner H, Amann R, Reinhardt R: Complete genome sequence of the marine planctomycete Pirellula sp. strain 1. PNAS. 2003, 100 (14): 8298-8303. 10.1073/pnas.1431443100.PubMed CentralView ArticlePubMedGoogle Scholar
- Sivakumar PM, Prabhawathi V, Doble M: Antibacterial activity and QSAR of chalcones against biofilm-producing bacteria isolated from marine waters. SAR QSAR Environ Res. 2010, 21 (3-4): 247-63. 10.1080/10629361003771009.View ArticlePubMedGoogle Scholar
- Cho JY, Kwon HC, Williams PG, Jensen PR, Fenical W: Azamerone, A Terpenoid Phthalazinone from a Marine-Derived Bacterium Related to the Genus Streptomyces. Actinomycetales. 2006, 8 (12): 2471-2474.Google Scholar
- Motohashi K, Irie K, Toda T, Matsuo Y, Kasai H, Sue M, Furihata K, Seto H: Studies on terpenoids produced by actinomycetes. 5-dimethylallylindole-3-carboxylic Acid and A80915G-8"-acid produced by marine-derived Streptomyces sp. MS239. J Antibiot (Tokyo). 2008, 61 (2): 75-80.View ArticleGoogle Scholar
- Paytan A, McLaughlin K: The Oceanic Phosphorus Cycle. Chem. Rev. 2007, 107: 563-576. 10.1021/cr0503613.View ArticlePubMedGoogle Scholar
- Mikhail VS, Fuchs BM, Eilers H, Burkill PH, Amann R: Determination of Total Protein Content of Bacterial Cells by SYPRO Staining and Flow Cytometry. Appl Env Microbiol. 1999, 65 (7): 3251-3257.Google Scholar
- Martin JH, Fitzwater SE: Iron-deficiency limits phytoplankton growth in the Northeast Pacific Subarctic. Nature. 1988, 331: 341-343. 10.1038/331341a0.View ArticleGoogle Scholar
- Martin JH, Coale KH, Johnson KS, Fitzwater SE, Gordon RM, Tanner SJ, Hunter CN, Elrod VA, Nowicki JL, Coley TL, Barber RT, Lindley S, Watson AJ, Van Scoy K, Law CS, Liddicoatp MI, Ling R, Stanton T, Stockel J, Collins C, Anderson A, Bidigare R, Ondrusek M, Latasa M, Millerostar FJ, Leestar K, Yao W, Zhangstar JZ, Friederich G, Sakamoto C, Chavez F, Buck K, Kolber Z, Greene R, Falkowski P, Chisholm SW, Hoge F, Swift R, Yungel J, Nightingale SP, Hatton A, Liss P, Tindale NW: Testing the iron hypothesis in ecosystems of the equatorial Pacific Ocean. Nature. 2002, 371: 123-129.View ArticleGoogle Scholar
- Martin JH: Glacial-interglacial CO2 change: The iron hypothesis. Paleoceanography. 1990, 5: 1-13. 10.1029/PA005i001p00001.View ArticleGoogle Scholar
- Street JH, Paytan A: Iron, phytoplankton growth, and the carbon cycle. Met Ions Biol Syst. 2005, 43: 153-93.PubMedGoogle Scholar
- Blain S, et al: Effect of natural iron fertilization on carbon sequestration in the Southern Ocean. Nature. 2007, 446: 1070-1074. 10.1038/nature05700.View ArticlePubMedGoogle Scholar
- Gilbert JA, Thomas S, Cooley NA, Kulakova A, Field D, Booth T, McGrath JW, Quinn JP, Joint I: Potential for Phosphonate utilisation by marine bacteria in temperate coastal waters. Environmental microbiology. 2009, 11 (1): 111-25. 10.1111/j.1462-2920.2008.01745.x.View ArticlePubMedGoogle Scholar
- Kulakova AN, Wisdom GB, Kulakov LA, Quinn JP: The purification and characterization of phosphonopyruvate hydrolase, a novel carbon-phosphorus bond cleavage enzyme from Variovorax sp Pal2. J Biol Chem. 2003, 278 (26): 23426-31. 10.1074/jbc.M301871200. Epub 2003 Apr 15View ArticlePubMedGoogle Scholar
- Quinn JP, Kulakova AN, Cooley NA, McGrath JW: New ways to break an old bond: the bacterial carbon-phosphorus hydrolases and their role in biogeochemical phosphorus cycling. Environmental Microbiology. 2007, 9 (10): 2392-2400. 10.1111/j.1462-2920.2007.01397.x.View ArticlePubMedGoogle Scholar
- Repeta DJ, Pantoja S, Johnson CG: Two chemically distinct pools of organic nitrogen accumulate in the ocean. Science. 2005, 308: 1007-1010. 10.1126/science.1108925.View ArticlePubMedGoogle Scholar
- Gooday GW: The ecology of chitin degradation. Edited by: Marshall KC. 1990, Advances in microbial ecology Plenum Press, Inc., New York, NY, 11: 387-429.Google Scholar
- Jeuniaux C, Vossfoucart MF: Chitin biomass and production in the marine environment. Biochem. Syst. Ecol. 1991, 19: 347-356. 10.1016/0305-1978(91)90051-Z.View ArticleGoogle Scholar
- Keyhani NO, Roseman S: Physiological aspects of chitin catabolism in marine bacteria. Biochim Biophys Acta. 1999, 1473 (1): 108-22.View ArticlePubMedGoogle Scholar
- Cottrell MT, Kirchman DL: Natural assemblages of marine proteobacteria and members of the Cytophaga-Flavobacter cluster consuming low- and high-molecular-weight dissolved organic matter. Appl Environ Microbiol. 2000, 66 (4): 1692-7. 10.1128/AEM.66.4.1692-1697.2000.PubMed CentralView ArticlePubMedGoogle Scholar
- Cottrell MT, Wood DN, Yu L, Kirchman DL: Selected chitinase genes in cultured and uncultured marine bacteria in the alpha- and gamma-subclasses of the proteobacteria. Appl Environ Microbiol. 2000, 66 (3): 1195-201. 10.1128/AEM.66.3.1195-1201.2000.PubMed CentralView ArticlePubMedGoogle Scholar
- Cottrell MT, Moore JA, Kirchman DL: Chitinases from uncultured marine microorganisms. Appl Environ Microbiol. 1999, 65 (6): 2553-7.PubMed CentralPubMedGoogle Scholar
- Svitil AL, Kirchman DL: A chitin-binding domain in a marine bacterial chitinase and other microbial chitinases: implications for the ecology and evolution of 1,4-beta-glycanases. Microbiology. 1998, 144 (Pt 5): 1299-308.View ArticlePubMedGoogle Scholar
- Southward AJ, Langmead O, Hardman-Mountford NJ, Aiken J, Boalch GT, Dando PR, Genner MJ, Joint I, Kendall MA, Halliday NC, Harris RP, Leaper R, Mieszkowska N, Pingree RD, Richardson AJ, Sims DW, Smith T, Walne AW, Hawkins SJ: Long-term oceanographic and ecological research in the Western English Channel. Adv Mar Biol. 2005, 47: 1-105.View ArticlePubMedGoogle Scholar
- Petković H, Cullum J, Hranueli D, Hunter IS, Perić-Concha N, Pigac J, Thamchaipenet A, Vujaklija D, Long PF: Genetics of Streptomyces rimosus, the oxytetracycline producer. Microbiol Mol Biol Rev. 2006, 70 (3): 704-28. 10.1128/MMBR.00004-06.PubMed CentralView ArticlePubMedGoogle Scholar
- Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, Amin N, Schwikowski B, Ideker T: Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 2003, 13 (11): 2498-504. 10.1101/gr.1239303.PubMed CentralView ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.