Open Access

Metagenomes and metatranscriptomes from the L4 long-term coastal monitoring station in the Western English Channel

  • Jack A. Gilbert1, 2, 3Email author,
  • Folker Meyer2, 3,
  • Lynn Schriml4,
  • Ian R. Joint1,
  • Martin Mühling5 and
  • Dawn Field6
Standards in Genomic Sciences20103:3020183

https://doi.org/10.4056/sigs.1202536

Published: 31 October 2010

Abstract

Both metagenomic data and metatranscriptomic data were collected from surface water (0–2m) of the L4 sampling station (50.2518 N, 4.2089 W), which is part of the Western Channel Observatory long-term coastal-marine monitoring station. We previously generated from this area a six-year time series of 16S rRNA V6 data, which demonstrated robust seasonal structure for the bacterial community, with diversity correlated with day length. Here we describe the features of these metagenomes and metatranscriptomes. We generated 8 metagenomes (4.5 million sequences, 1.9 Gbp, average read-length 350 bp) and 7 metatranscriptomes (392,632 putative mRNA-derived sequences, 159 Mbp, average read-length 272 bp) for eight time-points sampled in 2008. These time points represent three seasons (winter, spring, and summer) and include both day and night samples. These data demonstrate the major differences between genetic potential and actuality, whereby genomes follow general seasonal trends yet with surprisingly little change in the functional potential over time; transcripts tended to be far more structured by changes occurring between day and night.

Keywords

Marineaerobicsurface watercoastaltemperatemetagenomemetatranscriptomepyrosequencingtime-seriesdielseasonal

Introduction

The Western Channel Observatory station L4, located off the Plymouth coast in the UK, has been collecting environmental data for almost a century [1]. This includes published 16S rRNA V6 amplicon pyrosequencing data cataloging monthly patterns in microbial diversity [2,3]. The importance of the area rests with its being a transition zone between many northern and southern planktonic species [1] and with the fact that, as a major confluence between the North Atlantic Ocean and the North Sea, water masses exhibit extremely short residence times (>2 months [4];). In the study reported here, we use shotgun metagenomics and metatranscriptomics to identify the relationship between genetic and functional diversity at station L4.

Classification and features

Relationship of reported datasets

We generated 8 metagenomes and 7 metatranscriptomes for eight time points. Figure 1 shows the relationships of these metagenomes and metatranscriptomes; the figure was produced by using a group-average clustering dendrogram representing the relationships based on comparison of 66,529 amino acid sequences of greater than 40 amino acids predicted from each dataset (for details of the process, see Metagenome Annotation). One can clearly see that the metagenomic and metatranscriptomic data cluster separately. The metagenomic data shows an average similarity of less than 2%, clustered by season, from which one can infer that the seasonal differences are stronger than the diel differences. On the other hand, the metatranscriptomes show more similarity and a tendency to cluster by diel time point; specifically, the April night data and January night data are more similar to each other than either is to the April day data and January day data. The August metatranscriptomes cluster by themselves, but this clustering is also structured by day and night. Table 1 details the classification and general features of the metagenomic datasets information for this study in MIMS format.
Figure 1.

Group-average dendrogram showing relationship between all metagenomes and metatranscriptomes, based on comparison of annotated protein fragments via BLAST x using the SEED database in MG-RAST for each dataset. MTS – metatranscriptome. MGS – metagenome.

Table 1.

Classification and general feature of 8 metagenome datasets according to the MIMS recommendations [5].

MIGS ID

Property

Term

Evidence code

 

Current classification

Metagenome ecological metagenome marine metagenome

TAS [6]

5

Collection date

Jan Day: 2008-01-28T15:30

TAS [6]

Jan Night: 2008-01-28T19:00

Apr Day: 2008-04-22T16:00

Apr Night: 2008

Aug 4pm: 2008

Aug 10 pm: 2008

Aug 4 am: 2008

Aug 10 am: 2008

6

Latitude Longitude

Jan Day: 50.2518:4.2089

NAS

Jan Night: 50.2611:4.2435

Apr Day: 50.2518:4.2089

Apr Night: 50.2530:4.1875

Aug 4pm: 50.2518:4.2089

Aug 10 pm: 50.2545:4.1990

Aug 4 am: 50.2678:4.1990

Aug 10 am: 50.2665:4.1486

7

Depth

0

NAS

8

Altitude

0

NAS

9

Geographic location/Country

England

NAS

10

Environment

Coastal Marine

 

11a

Environmental Package

See Table 2

 

29

Sample collection device or method

Large bore peristaltic filtration pump

 

30

Sample material processing

Water filtered on to a 0.22 µm Sterivex (Millipore) filter and then snap-frozen at −80C

 

31

Amount or size of sample collected

10L

 

Evidence codes - IDA: Inferred from Direct Assay (first time in publication); TAS: Traceable Author Statement (i.e., a direct report exists in the literature); NAS: Non-traceable Author Statement (i.e., not directly observed for the living, isolated sample, but based on a generally accepted property for the species, or anecdotal evidence). These evidence codes are from the Gene Ontology project [7]. If the evidence code is IDA, then the property was directly observed for a live isolate by one of the authors or an expert mentioned in the acknowledgements.

Environmental characteristics and descriptions

Environmental data was collected for temperature, density, salinity, chlorophyll a, total concentration of organic nitrogen and carbon, nitrate, ammonia, silicate, and phosphate [Table 2]. The methods used are described on the Western Channel Observatory website.
Table 2.

Environmental variables for each sampling occasion

Property

Measurementa

Sample Collection date (MIGS-5)

01/28

01/28

04/22

04/22

08/26

08/26

08/27

08/27

Evidence code

Sample collection time

15:38

19:30

16:00

22:00

16:00

22.00

04:00

10:00

 

Temperature (°C)

10.1

10.1

9.7

9.6

15.9

15.8

15.7

15.8

IDA

Density (kg m−2)

1025.6

1026.3

1027.2

1027.1

1023.5

1024.3

1024.5

1024.4

 

Salinity (PSU)

33.3

34.2

35.1

35.0

32.1

33.0

33.3

33.2

 

Chlorophyll a (µg/L)

0.8

0.9

2.2

1.3

9.2

8.2

9.8

11.9

IDA

Total Organic Nitrogen (µmol L-1)

1.3

3.5

2.9

2.8

2.8

2.3

3.0

4.1

IDA

Total Organic Carbon (µmol L-1)

33.2

38.2

27.2

19.4

26.8

26.5

22.0

23.7

IDA

NO2 + NO3 (µmol L-1)

10.9

10.0

4.0

3.8

0.1

0.1

0.9

0.1

 

Ammonia (µmol L-1)

0.0

0.0

0.5

0.3

0.1

0.1

0.1

0.1

IDA

SRP (µmol L-1)

0.5

0.5

0.4

0.3

0.0

0.1

0.0

0.1

 

Silicate (µmol L-1)

6.0

5.8

2.6

2.7

0.1

0.2

0.3

0.2

 

a Samples collected January–August, 2008. Evidence codes: MIGS-5: TAS [5].

Figure 2 plots the environmental trends at L4 averaged for the years 2003–2008; the graph clearly shows the differences among the samples taken in the three months. Figure 3 shows a principal component analysis of the environmental parameters recorded during this study. Evident from the figure is the fact that the January samples have higher nutrient concentrations, the April samples show changes in the water salinity as a consequence of density, and the August samples show changes in temperature and chlorophyll a concentration.
Figure 2.

Monthly annual averages for all environmental parameters and species richness (S). TO — total organic; SRP — Soluble Reactive Phosphorous; PAR — Photosynthetically Active Radiation; NAO — North Atlantic Oscillation. Data taken from Gilbert et al., 2010.

Figure 3.

Principal component analysis of environmental variables showing the seasonal differences in variables outlined in Table 2. Classification and general features of the 15 datasets in accordance with the MIMS recommendations [5]

Metagenome sequencing and annotation

Metagenome project history

Two factors motivated the choice of station L4: its century-long history of environmental data [8] and the six years of 16S rRNA V6 amplicon pyrosequencing information detailing microbial diversity patterns [2,3], from which we inferred interannual variability from our single-year study. All 16S rRNA V6 amplicon pyrosequencing data have been submitted to the NCBI short reads archive under SRA009436 and registered with the GOLD database (Gm00104). The data also can be accessed from the VAMPS servers. The metagenomic data and metatranscriptomic data are available on the CAMERA website under Western Channel Observatory Microbial Metagenomic Study and on the Metagenome Rapid Annotation using Subsystem Technology (MG-RAST) system under 4443360-63, 4443365-68 and 4444077, 4445065-68, 4445070, 4445081, and 4444083, as well as through the INSDC short-reads archive under ERP000118. Table 1, Table 2, Table 3, and Table 4 detail the metagenomic sequencing project information for this study in MIMS format.
Table 3.

Metagenome sequencing project information (MIMS compliance)

MIGS ID

Property

Jan 3pm

Jan 7pm

Apr 4pm

Apr 10pm

Aug 4pm

Aug 10pm

Aug 4am

Aug 10am

35

library reads sequenced

616,793

784,823

637,801

493,003

620,759

524,953

500,117

326,475

32

nucleic acid extraction

Gilbert et al. 2008

43

sequencing method

454 Titanium pyrosequencing (GS flx)

46

Assembly

none

 

INSDC ID

SRA009436

 

GenBank Date of Release

01-12-2009

 

GOLD ID

GM00104

Table 4.

Metatranscriptome sequencing project information (MIMS compliance)

MIGS ID

Property

Jan 3pm

Jan 7pm

Apr 4pm

Apr 10pm

Aug 4pm

Aug 10pm

Aug 4am

35

library reads sequenced

139,880

130,826

124,925

147,492

139,375

193,254

154,865

32

nucleic acid extraction

Gilbert et al. 2008

43

sequencing method

454 Titanium pyrosequencing (GS flx)

46

Assembly

none

 

INSDC ID

SRA009436

 

GenBank Date of Release

01-12-2009

 

GOLD ID

GM00104

Sampling and DNA isolation

For the sampling, a minimal-impact surface buoy was deployed with a 7 m current drogue following a Lagrangian drift. Samples were taken at station L4 to represent three seasons and both day and night readings, as follows:
  • Winter: January 28, at 3:00 pm and again at 7 pm (2 hours after sundown) at 50.2611 N: 4.2435 W

  • Spring: April 22, at 4 pm and again at 10 pm (one and a half hours after sundown) at 50.253N:4.1875W

  • Summer: August 27, at 4 pm and again at 10 pm (two hours after sundown) at 50.2545N:4.199W

  • Summer: August 28, at 4 am (two hours before sunrise) at 50.2678N:4.1723W and at 10 am at 50.2665N:4.1486W

The sampling technique involved the following steps: (1) collection of 20 L of seawater from the surface (0–2 m), (2) prefiltering through a 1.6 µm GF/A filter (Whatmann), (3) passage of the filtrate through a 0.22 µm Sterivex cartridge (Millipore) for a maximum of 30 minutes (approximately 10 L per Sterivex cartridge); (4) pump-drying and snap-freezing of the cartridges in liquid nitrogen, (5) barcoding [9] of the samples at the laboratory, and (6) storage at −80 °C.

Both DNA and RNA then were isolated from each sample [2,9], barcoded, and stored at −80°C. DNA and mRNA-enriched cDNA were purified from the samples; for details, see [9].

Metagenome sequencing and assembly

The isolated DNA was used for metagenomic analysis, and the mRNA-enriched cDNA was used for metatranscriptomic pyrosequencing analysis. All DNA and cDNA were pyrosequenced on the GS-FLX Titanium platform. No DNA assembly was carried out.

Metagenome annotation

The MG-RAST bioinformatics server [10] was used for annotating the metagenomic samples [113]. The data also were processed by using custom-written programming scripts on the Bio-Linux system [6] at the NERC Environmental Bioinformatics Centre unless otherwise indicated. In order to ensure high quality, the following sequences were removed from the pyrosequenced data: transcript fragments with >10% non-determined base pairs (Ns), fragments <75 bp in length, fragments with >60% of any single base, and exact duplicates (resulting from aberrant dual reads during sequence analysis). So-called artificial duplicates in the metagenomic data (i.e., multiple reads that start at the same position; see, e.g., Gomez-Alvarez et al., 2009) were not removed, however, because of the possibility of their being natural; their removal would have precluded comparison with the metatranscriptomic data [12].

The nucleic acid sequences were then compared with three major ribosomal RNA databases – (SILVA, RDP II, and Greengenes – using the bacterial and archaeal 5S, 16S, and 23S and the eukaryotic 18S and 25S sequence annotator function of MG-RAST (e-value < 1 × 10–5; minimum length of alignment of 50 bp; minimum sequence nucleotide identity of 50%). Reads annotated as rRNA were excluded. All subsequent reads were considered to be valid DNA or valid putative mRNA derived sequences and 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%; Meyer et al., 2008). The result was an abundance matrix of functional genes and protein-derived predicted taxonomies across the DNA and mRNA samples.

The sequences also were translated using the techniques described by Gilbert et al. (2008) and Rusch et al. (2007) [9,13]. Predicted open reading frames (pORFs) having >40 amino acids were produced in all six reading frames. The CD-HIT program [15] was used to cluster the proteins from the datasets at 95% amino acid identity over 80% of the length of the longest sequence in a cluster. The longest representative from each cluster then was clustered at 60% amino acid identity over 80% of the length of the longest sequence to group these sequences by protein families. Based on the relative abundance of each sample in a cluster, an abundance matrix was created using the output cluster files from CD-HIT that contained the original fasta sequences and headers for each sample (abundanceMatrix-twoStep.pl). Subsequently, protein clusters with ≤2 representative pORFs were removed from the pORF matrix (MatrixParser.pv). In order to equalize the sequencing effort, all samples were randomly resampled (Daisychopper.pl) to the same number of pORFs or sequences across the clusters or functional/taxonomic SEED annotations.

Metagenome properties

Approximately 4.5 million combined microbial metagenomic reads were produced, comprising 1.9 billion bp, with an average read length of 350 bp across the eight samples, ranging from 326,475 to 784,823 sequences [Table 5]. Seven metatranscriptomic datasets were also produced (the sample taken on August 28 at 10 am was lost in transit) totaling 1 million sequences. After cleanup, 392,632 putative mRNA-derived sequences remained, totaling 159 million bp, with an average of 272 bp per sequence. The effort per sample varied from 33,149 to 96,026 sequences [Table 6]. SEED annotations produced via MG-RAST (Table 7 and Table 8 ranged from 20% to 46% of each metagenomic dataset and from to 11% to 35% of the metatranscriptomic datasets.
Table 5.

Metagenome statistics

 

Jan 3pm

Jan 7pm

Apr 4pm

Apr 10pm

Aug 27 4pm

Aug 27 10pm

Aug 28 4am

Aug 28 10am

No. Original DNA Sequences

616,793

784,823

637,801

493,003

620,759

524,953

500,117

326,475

Predicted ORFs (>40aa pORFs)

862,695

1,287,412

1,003,799

745,305

986,269

846,209

779,951

491,330

No. of pORF clusters (95%)

615,374

1,123,829

779,342

588,387

881,113

703,712

675,210

444,729

No. of pORF singletons (95%)

546,463

1,031,865

682,586

526,233

805,284

634,042

608,785

410,616

No. of pORF ‘families’ (60%)

423,674

1,031,904

678,547

528,213

801,760

637,542

620,403

419,461

No. of pORF singletons (60%)

352,938

962,073

609,351

486,712

740,032

589,839

577,027

398,202

Resampled pORFs (66529)

        

No. of pORF clusters (95%) (66529)

56337

64446

61187

59904

65601

63032

64729

65075

No. of pORF singletons (95%) (66529)

52891

63378

58691

57779

64818

61068

63359

63945

Good’s Coverage (66529)

20.50

4.74

11.78

13.15

2.57

8.21

4.76

3.88

No. DNA seqs with functional annotation

122,936

291,953

258,658

164,249

283,761

196,369

196,972

126,392

No. DNA seqs without functional annotation (%)

493,857

492,870

379,143

328,754

336,998

328,584

303,145

200,083

Percent DNA seqs without functional annotation

80%

63%

59%

67%

54%

63%

61%

61%

No. DNA seqs with taxonomic annotation

190,326

417,920

349,888

241,541

379,911

288,356

304,003

186,421

Resampled sequencing effort (186,421)

        

Number of archaeal sequences (186,421)

19,055

15,150

777

561

1,370

1,093

1,585

1,244

Number of bacterial sequences (186,421)

161,899

146,911

182,850

180,674

182,717

176,825

180,725

182,332

Table 6.

Metatranscriptome statistics

 

Jan 3pm

Jan 7pm

Apr 4pm

Apr 10pm

Aug 27 4pm

Aug 27 10pm

Aug 28 4am

No. Original cDNA Sequences

139,880

130,826

124,925

147,492

139,375

193,254

154,865

No. of sequences following filtering***

94,024

106,864

84,916

109,577

87,799

118,360

111,568

No. mRNA following removal of rRNA

61,831

96,026

41,378

53,413

33,149

51,829

55,006

Predicted ORFs (>40aa pORFs)

143,169

211,374

81,642

107,699

77,985

66,529

159,909

No. of pORF clusters (95%)

98,871

78,278

35,648

51,088

28,167

24,136

68,080

No. of pORF singletons (95%)

82,464

54,870

25,925

38,960

19,600

17,177

50,246

No. of pORF ‘families’ (60%)

84,598

45,049

19,131

37,628

15,146

12,735

41,480

No. of pORF singletons (60%)

76,655

30,720

13,869

30,919

9,857

9,134

32,662

Resampled pORFs (66529)

       

No. of pORF clusters (95%) (66529)

31026

50354

30334

34217

24848

24136

33191

No. of pORF singletons (95%) (66529)

23038

43687

22394

26840

17373

17177

25636

Good’s Coverage (66529)

65.37

34.33

66.34

59.66

73.89

74.18

61.47

No. mRNA seqs with functional annotation

11,513

31,990

8,845

16,315

11,720

5,907

15,384

No. mRNA seqs without functional annotation

50,318

64,036

32,533

37,098

21,429

45,922

39,622

Percent DNA seqs without functional annotation

81%

67%

79%

69%

65%

89%

72%

No. mRNA seqs with taxonomic annotation

29,521

30,778

20,899

26,398

15,456

29,605

38,304

Resampled sequencing effort (15,456)

       

Number of archaeal sequences (15,456)

625

49

1

16

4

4

11

Number of bacterial sequences (15,456)

13,633

11,926

13,702

8,449

14,469

15,071

14,803

Table 7.

Number of genes associated with the general SEED functional categories

Subsystem Hierarchy 1

Jan 3pm

Jan 7pm

April 4pm

April 10pm

Aug 27 4pm

Aug 27 10pm

Aug 28 4am

Aug 28 10am

Amino Acids and Derivatives

13,515

12,346

13,913

12,089

13,279

12,517

11,966

12,074

Carbohydrates

14,181

13,087

14,884

13,829

14,801

13,929

13,258

13,780

Cell Division and Cell Cycle

2,136

2,026

2,286

2,243

2,243

2,231

2,175

2,234

Cell Wall and Capsule

5,632

5,363

5,336

6,051

5,553

5,674

6,079

6,347

Clustering-based subsystems

18,051

17,585

19,425

19,647

19,055

19,441

20,434

19,860

Cofactors, Vitamins, Prosthetic Groups, Pigments

8,497

7,675

8,188

8,606

8,142

8,227

8,582

8,001

DNA Metabolism

5,461

5,331

5,191

5,559

5,321

5,717

5,824

5,855

Fatty Acids and Lipids

2,165

1,919

1,883

1,891

1,955

2,025

1,960

1,934

Macromolecular Synthesis

148

147

287

163

213

151

136

109

Membrane Transport

2,764

2,322

2,839

2,375

2,606

2,507

2,234

2,234

Metabolism of Aromatic Compounds

1,817

1,357

1,473

1,527

1,632

1,409

1,629

1,489

Miscellaneous

381

367

448

423

417

446

454

393

Motility and Chemotaxis

1,034

994

879

1,227

977

1,203

1,311

1,348

Nitrogen Metabolism

668

688

587

574

747

718

628

660

Nucleosides and Nucleotides

5,152

4,820

4,701

4,578

4,836

4,752

4,639

4,706

Phosphorus Metabolism

1,796

1,706

1,747

1,926

1,832

1,958

2,085

1,879

Photosynthesis

212

4,373

160

1,489

127

197

270

203

Potassium metabolism

648

591

586

631

620

755

838

817

Protein Metabolism

11,912

11,717

11,254

11,534

11,473

11,597

11,210

11,715

RNA Metabolism

5,133

4,889

4,660

4,813

4,811

4,744

5,068

4,981

Regulation and Cell signaling

1,196

1,127

1,400

966

1,356

1,360

1,076

1,056

Respiration

5,298

8,480

5,455

5,570

5,432

5,579

4,926

4,994

Secondary Metabolism

116

124

63

87

93

83

86

83

Stress Response

2,497

2,133

2,338

2,419

2,306

2,524

2,508

2,605

Sulfur Metabolism

1,604

1,354

1,673

1,430

1,446

1,240

1,320

1,317

Unclassified

6,235

5,677

6,567

5,763

6,672

6,019

5,555

5,794

Virulence

4,686

4,733

4,711

5,521

4,989

5,929

6,684

6,467

Table 8.

Number of transcripts associated with the general SEED functional categories

Subsystem Hierarchy 1

Jan 3:30pm

Jan 7pm

April 4pm

April 10pm

Aug 27 4pm

Aug 27 10pm

Aug 28 4am

Amino Acids and Derivatives

261

536

204

198

21

144

443

Carbohydrates

886

1767

546

1302

530

1381

1256

Cell Division and Cell Cycle

83

191

52

63

96

56

80

Cell Wall and Capsule

154

353

317

297

153

113

221

Clustering-based subsystems

641

657

294

451

111

157

427

Cofactors, Vitamins, Prosthetic Groups, Pigments

215

457

130

248

24

13

469

DNA Metabolism

102

108

83

122

24

26

85

Fatty Acids and Lipids

84

28

17

27

0

28

10

Macromolecular Synthesis

0

0

5

2

2

0

0

Membrane Transport

44

19

237

83

2673

13

440

Metabolism of Aromatic Compounds

47

6

16

4

0

24

14

Miscellaneous

53

80

54

55

672

43

75

Motility and Chemotaxis

40

10

438

58

3

8

180

Nitrogen Metabolism

11

0

0

2

9

8

3

Nucleosides and Nucleotides

144

87

42

48

4

13

56

Phosphorus Metabolism

79

83

64

94

25

18

31

Photosynthesis

67

0

17

2

0

1

0

Potassium metabolism

29

13

3

13

4

2

7

Protein Metabolism

439

95

129

625

81

112

172

RNA Metabolism

1631

160

1813

702

907

2883

874

Regulation and Cell signaling

65

136

16

354

30

18

41

Respiration

174

20

26

97

125

31

109

Secondary Metabolism

18

3

1

0

0

0

1

Stress Response

100

175

42

229

5

43

56

Sulfur Metabolism

42

18

19

14

13

11

40

Unclassified

346

58

957

101

10

110

271

Virulence

152

847

385

716

385

651

546

Highlights from the metagenome sequences

In general, in the samples, metagenomes were more similar than metatranscriptomes. Photosynthesis genes showed both seasonal and diel changes: specifically, 10 times greater photosynthetic potential in winter than in summer and greater abundance at night in January and April. Gene fragments annotated to proteorhodopsin showed virtually no seasonal or diel fluctuations, however: only approximately 0.07% of the annotated functional profile from each sample. Other seasonal differences in metagenomic profiles included a considerably higher winter abundance (compared to spring or summer) of archaeal genes associated with lipid synthesis, thermosome chaperonins, RNA polymerase, small subunit ribosomal proteins, DNA replication, and rRNA modification. Diel differences were apparent among genes involved in respiratory metabolism, which were more abundant at night.

The metatranscriptomic photosynthetic profiles were similar to those of the metagenomes in that photosynthesis genes were most abundant in January and virtually absent in August. Photosynthetic transcripts also were most abundant during the winter. On the other hand, unlike metagenomes, they were most abundant in the daytime in all months. Other seasonal differences in metatranscriptomic seasonal profiles included a greater abundance of transcripts related to membrane transport, especially amino acid transport, in summer when nutrients and dissolved organic material (DOM) are least abundant. The diel metatranscriptional profiles for January showed considerable difference in functions (in addition to photosynthesis); for example, transcripts relating to nitrogen cycling were most abundant during the day and were associated mainly with ammonification. Cell wall and capsule and cell division and cycle were upregulated at night, suggesting a nocturnal increase in cell division, potentially associated with the Cyanobacteria. Similarly, April samples showed a considerable up-regulation in RNA metabolism during the day, resulting primarily from an increase in group I intron and RNA polymerase transcripts. In August, transcripts with homology to membrane transport were upregulated during the day, while transcripts associated with motility and chemotaxis and with the synthesis of cofactors, vitamins, prosthetic groups, and pigments were considerably upregulated at night, suggesting that nocturnal motility and cellular activity (nucleotide and amino acid synthesis) were also upregulated.

Declarations

Acknowledgements

This work was funded by a grant from the Natural Environmental Research Council (NERC - NE/F00138X/1). The authors thank Neil Hall from the NERC / University of Liverpool Advanced Genomics Facility. This work was supported in part by the Office of Advanced Scientific Computing Research, Office of Science, U.S. Department of Energy, under Contract DE-AC02-06CH11357. The submitted manuscript has been created by UChicago Argonne, LLC, Operator of Argonne National Laboratory (“Argonne”). Argonne, a U.S. Department of Energy Office of Science laboratory, is operated under Contract No. DE-AC02-06CH11357. The U.S. Government retains for itself, and others acting on its behalf, a paid-up nonexclusive, irrevocable worldwide license in said article to reproduce, prepare derivative works, distribute copies to the public, and perform publicly and display publicly, by or on behalf of the Government.

Authors’ Affiliations

(1)
Plymouth Marine Laboratory
(2)
Argonne National laboratory
(3)
University of Chicago
(4)
University of Maryland School of Medicine
(5)
IÖZ - Interdisciplinary Centre for Ecology, TU Bergakademie Freiberg
(6)
NERC Centre for Ecology and Hydrology

References

  1. Southward AJ, Langmead O, Hardman-Mountford NJ, Aiken J, Boalch GT, Dando PR, Genner MJ, Joint I, Kendall MA, Halliday NC, et al. Longterm oceanographic and ecological research in the Western English Channel. Adv Mar Biol 2004; 47:1–105. doi:10.1016/S0065-2881(04)47001-1View ArticleGoogle Scholar
  2. 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. Environ Microbiol 2009; 11:3132–3139. PubMed doi:10.1111/j.1462-2920.2009.02017.xView ArticlePubMedGoogle Scholar
  3. Gilbert JA, Swift P, Somerfield P, Temperton B, Huse S, Smyth T, Field D. Seasonal succession and impact of environmental change on bacterial populations in the Western English Channel: A sixyear study. ISME J 2010; (In Review).Google Scholar
  4. Siddorn JR, Allen JI, Uncles RJ. Heat, alt and tracer transport in the Plymouth Sound coastal region: a 3-D modeling study. J Mar Biol Assoc UK 2003; 83:673–682. doi:10.1017/S002531540300763XhView ArticleGoogle Scholar
  5. Field D, Garrity G, Gray T, Morrison N, Selengut J, Sterk P, Tatusova T, Thomson N, Allen MJ, Angiuoli SV, et al. The minimum information about a genome sequence (MIGS) specification. Nat Biotechnol 2008; 26:541–547. PubMed doi:10.1038/nbt1360PubMed CentralView ArticlePubMedGoogle Scholar
  6. Field D, Tiwari B, Booth T, Houten S, Swan D, Bertrand N, Thurston M. Open software for biologists: from famine to feast. Nat Biotechnol 2006; 24:801–803. PubMed doi:10.1038/nbt0706-801View ArticlePubMedGoogle Scholar
  7. Harris R. The L4 time-series: the first 20 years. J Plankton Res 2010; 32:577–583. doi:10.1093/plankt/fbq021View ArticleGoogle Scholar
  8. Booth T, Gilbert JA, Neufeld JD, Ball J, Thurston M, Chipman K, Joint I, Field D. Handlebar: a flexible, web-based inventory manager for handling barcoded samples. Biotechniques 2007; 42:300–302. PubMed doi:10.2144/000112385View ArticlePubMedGoogle Scholar
  9. Gilbert JA, Field D, Huang Y, Edwards R, Li W, Gilna P, Joint I. Detection of large numbers of novel sequences in the metatranscriptomes of complex marine microbial communities. PLoS ONE 2008; 3:e3042. PubMed doi:10.1371/journal.pone.0003042PubMed CentralView ArticlePubMedGoogle Scholar
  10. Meyer F, Paarmann D, D’Souza M, Olson R, Glass EM, Kubal M, Paczian T, 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. PubMed doi:10.1186/1471-2105-9-386PubMed CentralView ArticlePubMedGoogle Scholar
  11. Gomez-Alvarez V, Teal TK, Schmidt TM. Systematic artifacts in metagenomes from complex microbial communitiesMetagenomes artifact. ISME J 2009; 3:1314–1317. PubMed doi:10.1038/ismej.2009.72View ArticlePubMedGoogle Scholar
  12. Niu B, Fu L, Sun S, Li W. Artificial and natural duplicates in pyrosequencing reads of metagenomic data. BMC Bioinformatics 2010; 11:187. PubMed doi:10.1186/1471-2105-11-187PubMed CentralView ArticlePubMedGoogle Scholar
  13. Rusch DB, Halpern AL, Sutton G, Heidelberg KB, Williamson S, Yooseph S, Wu D, Eisen JA, Hoffman JM, Remington K, et al. The Sorcerer II Global Ocean Sampling Expedition: Northwest Atlantic through Eastern Tropical Pacific. PLoS Biol 2007; 5:e77. PubMed doi:10.1371/journal.pbio.0050077PubMed CentralView ArticlePubMedGoogle Scholar
  14. Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, Davis AP, Dolinski K, Dwight SS, Eppig JT, et al. Gene Ontology: tool for the unification of biology. Nat Genet 2000; 25:25–29. PubMed doi:10.1038/75556PubMed CentralView ArticlePubMedGoogle Scholar
  15. Li W, Godzik A. Cd-hit: a fast program for clustering and comparing large sets of protein or nucleotide sequences. Bioinformatics 2006; 22:1658–1659. PubMed doi:10.1093/bioinformatics/btl158View ArticlePubMedGoogle Scholar

Copyright

© The Author(s) 2010