High throughput sequencing to assess microbial diversity in four western hemispheric soils.
Luiz F.W. Roesch1,7, Roberta R. Fulthorpe2, Alberto Riva3, George Casella4, Angela D. Kent5,
Samira Daroub6, Flávio A. O. Camargo7, William G. Farmerie8, and Eric W. Triplett1
1
Department of Microbiology and Cell Science, Institute for Food and Agricultural Sciences,
University of Florida, Gainesville, FL 32611-0700 USA
2
Department of Physical and Environmental Sciences, University of Toronto Scarborough,
Toronto, Ontario, M1C 1A4 Canada
3
Department of Molecular Genetics and Microbiology, College of Medicine, Gainesville, FL
32610-0266 USA
4
Department of Statistics, University of Florida, Gainesville, FL 32611-8545 USA
5
Department of Natural Resources and Environmental Sciences, University of Illinois at UrbanaChampaign, Urbana, IL 61801 USA
6
Department of Soil and Water Science, Everglades Research and Education Center, Institute for
Food and Agricultural Sciences, University of Florida, Belle Glade, FL 33430 USA
7
Department of Soil Science, Federal University of Rio Grande do Sul, Porto Alegre, Brazil
8
Interdisciplinary Center for Biotechnology Research, University of Florida, Gainesville, FL
32610 USA
ABSTRACT
The objective of this work was to estimate the number of operational taxonomic units
(OTUs) in a gram of soil using 454 pyrosequencing. Soil samples from four sites across a large
transect of the western hemisphere were collected, DNA isolated, and the 3’-end of the 16S
rRNA gene amplified. Sites included a boreal forest soil from northwestern Ontario, a temperate
zone agricultural soil from the Morrow Plots in Illinois, a tropical, highly organic soil from the
Everglades Agricultural Area in Florida, and a temperate agricultural soil from southern Brazil.
Pyrosequencing was performed so that only the most hypervariable region (approximately 13901500) of the 16S rRNA gene was sequenced. The sequences were aligned using MUSCLE,
distance matrices determine by DNADist, and clutering done by DOTUR. The output from
DOTUR was the theoretical maximum OTU number determined by the Transform Both Sides
Power of x (TBS-PX) transformation of Michaelis-Menton kinetics. The number of sequences
available for each site varied from 26,140 to 53,533. Crenarchaeotal sequences were also
amplified with far more obtained from the agricultural soils than the forest soil. The most
abundant groups in all four soils appear to the Bacteriodetes and the - and -Proteobacteria.
The number of OTUs was calculated using parametric and non-parametric estimators and at five
levels of dissimilarlity in nucleotide sequence. Even at the lowest level of dissimilarity using the
most generous estimator, the number of OTUs in a gram of soil never exceeded 35,000,
considerably less than the 8.3 million estimated by Gans et al. (2005). The bacterial diversity of
the forest soil was strikingly different that those of the three agricultural soils. The forest soil
appears to be phylum rich and relatively species poor compared to the agricultural soils which
appeared to be species rich but phylum poor.
INTRODUCTION
Recently, much attention has been paid toward estimating the number of species or
operational taxonomic units (OTUs) in a gram of soil. This is of particular interest as soil is
considered to harbor the most diverse populations of bacteria of any environment on earth. Even
th0se soils with large particle sizes provide an enormous surface area for many bacteria. The
number of potential microhabitats within a small soil sample is extraordinary.
In the first culture-independent analysis of a bacterial species census in soil, Torsvik et al.
(1990) isolated DNA from cultured species and estimated the number of bacterial genomes
present in a mixed sample using DNA:DNA hybridization. Taking the admittedly risky
approach of extrapolating such data from a mixture containing a few species to a complex
mixture of DNA from soil, Torsvik et al. (1990) estimated that the number of bacterial species in
a gram of boreal forest soil was approximately 10,000. Gans et al. (2005) re-evaluated this
approach and predicted nearly 107 microbial species per gram of soil and proposed that this
estimate could be practically verified with current DNA sequencing technology.
In the meantime, others have attempted to estimate the number of bacterial species in a
gram of soil by extrapolation based on a sample size of amplified, cloned, and sequenced 16S
rRNA genes. Schloss and Handelsman (2006) defined an operational taxonomic unit as 3%
dissimilarlity in the comparison of 16S rRNA gene sequences. Using that definition and two
sequenced clone libraries from different sites, they estimated the number of OTUs in a gram of
soil \to be between 2,000 and 5.000. These two libraries were derived from Alaskan and
Minnesota soils and contained 1,033 and 600 16S rRNA clones, respectively. Hong et al. (2006)
used a nonparametric approach in the analysis of 556 16S rRNA clones and estimated the
number of species per gram of marine sediment to be between 2,000 and 3,000.
None of the estimates to date used a sufficient number of sequences to reach the limit of a
rarefaction curve depicting the number of species versus the number of clones sequenced. In this
work, we used pyrosequencing to obtain well over 25,000 16S rRNA gene fragments from each
of four soils. This is an order of magnitude higher than done previously. A statistical
transformation of the resulting rarefaction curves and species richness indicators permits an
estimate of the number of OTUs, as defined by specific levels of sequence identity, for each soil.
The goal of this work was to get an estimate of soil bacterial diversity based on a number
of sequences that exceeds or approaches the estimates to date on the number of species in a gram
of soil. To do this, soil DNA was isolated and 454 pyrosequencing was employed following the
amplification of a hypervariable region of the highly conserved 16S rRNA gene.
Another objective was to determine how this estimate might vary across a broad
geographical scale. Soils from four disparate sites across the western hemisphere were chosen.
Three of these soils were agricultural soils. One agricultural soil was selected from a maize field
in the southern most state of Brazil, Rio Grande do Sul. Another agricultural soil came from a
sugarcane field in the Everglades Agricultural Area in Florida where organic matter content is
very high. A third agricultural soil came from the Morrow Plots at the University of Illinois in
Urbana. The fourth soil was derived from a boreal forest site in northwestern Ontario, Canada.
METHODS
Soil sampling sites and methods
From each site, soil was collected from the top 10 cm from the surface. For the boreal
forest soil. sphagnum and duff layers were removed prior to collection. All soil was packed on
ice upon collection and sent to Gainesville, Florida by express mail for futher analysis. Soils
were stored in the lab at -80°C until extraction.
At the time of collection, the soil from the Morrow Plots in Illinois was in a corn-oatsalfalfa rotation for 53 years and has received fertilization with limestone, nitrogen, phosphorous,
and potassium. The Everglades soil in Florida was planted with sugarcane and the Brazilian plot
was at the end of a maize growing cycle.
DNA Extraction and 16S rRNA gene amplification
Bacterial DNA was isolated from soil samples using the FastDNA® Kit (Qbiogene, Inc.,
Calif.) following the manufacturer’s instructions. The DNA obtained was cleaned from organic
and inorganic compounds that contaminate DNA soil preparations using the agarose-embedded
genomic DNA method described by Moreira (1998).
To amplify 16S rRNA gene sequences from bacteria, the conserved areas of sequences in
the ribosomal database were examined to determine the variability and length of each of the V1V9 variable regions in the gene. This was necessary since the read lengths obtained from GS-20
454 pyrosequencing is only about 105 bases on average. In addition, we needed to identify a
region of about 700-800 bases to amplify since 454 library construction is optimal with a
fragment of this size. One end of this region must flank a highly variable region that will be
sequenced to give us the maximum amount of phylogenetic information possible.
The analysis included aligning representative sequences from each phyla using ClustalW
and showed that the V9 region was optimal in length and variability for this purpose. To amplify
a fragment of the appropriate size, primers 780f and a modification of 1492r were chosen. The
1492r primer has been heavily used in 16S rRNA gene amplifications since its publication in
1991 (Lane 1991). However, the original 1492r sequence f0 (5'-GGTTACCTTGTTACGACTT3’) is not conserved at all sites. The first G and first T are variable according to alignment of
sequences from a newer database. Therefore, a modified primer, 1492-rm, was used (5'GNTACCTTGTTACGACTT-3'), where the 3' end corresponds to the 1492 position of E. coli
16S rRNA gene, but the primer is one bp shorter. This primer sequence was found in 36,850 of
262,030 of the bacterial sequences in RDP II, as compared to only 10,860 of 262,030 for the
original 1492r primer. These numbers are low overall since not many of the published sequences
have data at this end.
The modified primer, 1492-rm, gave the highest number of hits for this region of any of
the modified 1492 primers examined. For the other end of the area 787f was chosen, just prior to
the V5 region. This primer (5'-AATAGATACCCNGGTAG-3’) matched 157,088 of 262,030
sequences in RDP II, the high number reflects the greater number of sequences from this region
in the database.
Assuming 1 x 109 cells/g of soil, an average bacterial genome size of 3 MB, and an
average 16S rRNA copy number of one per cell, approximately 20 PCR cycles was required to
obtain 3 g of amplified 16S rRNA gene product with 96 PCR reactions each having 10 g of
template DNA. Three g of amplified 16S rRNA gene product from each soil was required to
construct the four libraries for 454 sequencing. The PCR conditions used were 94°C for 2
minutes, 20 cycles of 94°C for 45 s denaturation, 55°C for 45 s annealing, and 72°C for 1 min
extension followed by 72°C for 6 min.
After the first 20 rounds of amplification, another three rounds of amplifcation was done
to add the A and B adapters required for 454 pyrosequencing (Margulies et al. 2005) to specific
ends of the amplified 16S rRNA fragment for library construction. For this purpose, a new
primer was synthesized where the A adaptor sequence was immediately upstream of the 1492-rm
primer
sequence.
The
resulting
sequence
was
5'(CCATCTCATCCCTGCGTGTCCCATCTGTTCCCTCCCTGTCTCAG)GITACCTTGTTACG
ACTT-3' with the A adaptor sequence in parentheses. Similarly, the biotinylated B-adaptor
sequence
was
added
upstream
of
the
780f
primer
resulting
in
5’(BioTEG/CCTATCCCCTGTGTGCCTTGCCTATCCCCTGTTGCGTGTCTCAG)ATTAGATA
CCCIGGTAG-3' with the adaptor sequence in parentheses.
The resulting PCR product was nebulized and used in the emPCR process necessary to
make the single strands on beads as required for 454 pyrosequencing (Margulies et al. 2005).
Sequencing was done by separating the microtiter plate into four sections so that all four samples
could be sequenced in a single 454 run. The number of reads per samples, the maximum and
average read lengths per sample, and GenBank accession numbers are shown in Table 1.
Phylogenetic assignment
the 16S rRNA gene fragments were phylogenetically assigned according to their best
matches to sequences in the NAST (Nearest Alignment Space termination) database (DeSantis et
al. 2006).
The sequences were submitted to a web-interface tool
(http://greengenes.lbl.gov/NAST) which compares the sequences with a 16S rRNA reference
database with ~100,000 aligned sequences. Each sequence was aligned and categorized
according to the best match in the reference database. Rarefaction curves, richness estimators
and diversity indexes were calculated based on best matches for each query sequence to
sequences in NAST and their frequency of recovery. Phylogenic inferences were made according
to the closest sequences assigned in the NAST database.
Alignment and clustering of 16S rRNA gene fragments
A second approach was to cluster sequences into groups of defined sequences by using
DOTUR (Schloss and Handelsman, 2005). Multiple sequence alignment was done using
MUSCLE (Edgar, 2004) (with parameter -maxiters 1, -diags1 and -sv). Based on the alignment,
a distance matrix was constructed using DNAdist from PHYLIP suite of programs version 3.6
with default parameters (Felsenstein 1989, 2005). These pairwise distances served as input to
DOTUR (Schloss and Handelsman, 2005) for clustering the sequences into OTUs of defined
sequence identity that ranged from unique sequences (no variation) to 20% dissimilarity. These
clusters served as OTUs for generating rarefaction curves and for making calculations with the
species richness, sequence coverage, and species diversity indexes. To obtain ACE, Chao1, and
log normal estimators, the distance matrices were further analysed by DOTUR (Schloss and
Handelsman 2005). The capacity of each of these programs were exceeded by these datasets on
traditional laboratory PCs. Hence, these programs were run on an HP DL585 Proliant Server
with 64 Gigabytes of RAM and 2 dual core Opteron processors at 1.8GHz, running the RedHat
Enterprise Linux OS.
Statistical analysis of taxa richness - Rarefaction
At each percent and sample size, a Michaelis-Menten equation was fit to the dotur curve.
(For example, for the Brazil data, the dotur output was obtained for sample sizes
5000,10000,15000,20000,25000,26140, and percents 0,.03.,.05,.1 ,.2. The MM equation has
parameters V and K with the equation y=Vx/(K+x), and the estimate of V is our estimate of the
asymptote.
We found that the Michaelis-Menton equation was not a good fit to the rarefaction curve,
as two parameters were not able to sufficiently model the curvature. The resulting asymptote
estimates were actually below the observed data by a substantial amount. The problem was that
the curvature at the beginning of the rarefaction curve was different from that at the end. To
solve this problem we used a model of two Michaelis-Menton equations, one for the beginning
and one for the end of the curve, that were joined continuously at a point that was optimized by
the fit.
The curves were fit by the method of maximum likelihood, using the transform-bothsides methods of Rupert et al. (1989). This approach better models the error terms and provides
improved estimators over standard approaches such as Lineweaver-Burk. The result of this
analysis was an estimate of the asymptote for each percent.
Statistical analysis of taxa richness – Chao1, Ace and Rarefaction ourput
For each region we have an asymptote estimate for each sample size. To this curve (see
picture) we fit an MM equation to estimate the overall asymptote. (Here we typically have 5-7
data points, and the ordinary MM, with the TBS fit, works fine). We used the bootstrap to
estimate standard deviations for both V and K for each curve. We can also translate that in
standard deviations for estimating proportions of V. For example, if you want to know the
necessary sample for getting p*V species, where p is (say) .9, the answer is sample size (p/(1p))*K with standard error (p/(1-p))*Std(K).
Note that at low percents, the estimates are quite variable, but still in the same ballpark.
For higher percents, the estimates become remarkably similar.
RESULTS AND DISCUSSION
Distribution of divisions identified
With over 40% of the total bacterial sequences, the proteobacteria represented the
dominant division in each soil. The -proteobacteria were the dominant subdivision among the
proteobacteria in all soils except Brazil where the -proteobacteria were dominant. With 15-25%
of the bacterial sequences in each sample, the second most abundant division in all four soils was
the Bacteriodetes. Other prominent divisions were the Acidobacteria, the Actinobacteria, and the
Firmicutes. Among the other divisions, the Gemmatimonadetes represented 3.5% of the
sequences in the Canadian sample while in Florida and Brazil, the Nitrospira represented 3 and
2% of total sequences, respectively. In Illinois and Canada, the Verrucomicrobia were
represented by more than 2% of the sequences. In Illinois, nearly 4% of the sequences were in
the TM7 division. Approximately 6-12% of the sequences from each sample remained
unclassified.
Crenarchaeota diversity
The primers used in this work also amplified the crenarchaeota (Table 1). In the
agricultural soils, the crenarchaeota represented a fairly significant proportion, 4.6-12.5%, of the
total sequences. The surprising result was the relative dearth of crenarchaeotal sequences in the
Canadian sample with just five of 53,251 sequences being in this group. These five sequences
were all related to the ammonia oxidizing archaea and the low number of these sequences may
be attributable to the low pH of this soil. However, the Brazilian soil had an even lower pH and
yet over 4% of the total sequences from Brazil were crenarchaeotal. Further work is needed to
confirm this observation of a low proportion of crenarchaeotal sequences in the Canadian forest
soil. The cause of this may be that nitrogen is more limiting in forest soils compared to fertilized
agricultural soils.
Of the crenarchaeota observed from each of the agricultural soils, about one-third of the
sequences are closely related to the ammonia oxdidizing archaea. While 16S rRNA sequence
alone cannot describe the physiology of an organism it is no unreasonable to expect a high
number of ammonia oxidizing archaea in these soils given recent work (Leininger et al.
2006;Treusch et al. 2005).
Estimation of the number of OTUs in each soil
With an average read length of 105 bases in each fragment, it is impossible to define
species even with the hypervariable region sequenced. However, even the entire length of the
16S rRNA gene cannot be used in a species definition. Genera can be defined at approximately
the 5% level of dissimilarity. Five levels of dissimilarity are presented here for each sample so
that the reader can choose a level of discrimination of interest. Three estimates of the number of
OTUs at each % dissimilarity are presented. One of these estimates is parametric (an estimate
based on rarefaction) while two of these are based on non-parametric measures (Chao1 and Ace).
Using these estimates, the number of sequences required to reach 90 and 95% of the maximum
number of OTUs at each % dissimilarity are presented.
Even at 0% dissimilarity, the maximum number of OTUs in any soil using parametric or
non-parametric estimates, is less than 35,000. This is over 200-fold lower than the estimate of
OTUs suggested by Gans et al. (2005) in a gram of soil. The maximum number of sequences
required to identify the 35,000 OTUs is never more than 600,000. Thus, using the methods here,
95% of all OTUs can be identified with over 10-fold fewer sequences than the number of species
suggested by Gans et al. (2005). As a result, we strongly disagree with the statement of Gans et
al. (2005) that “the survey size required for accurate analysis of soil communities is impractically
large”. In fact, with the latest improvement in throughput in 454 pyrosequencing, this survey can
be completed in less than one day of operation of the 454 sequencer.
We are aware of the biases involved in our methods. We are aware that our soil
extraction method likely did no complete extract all DNA present in the soil. We estimate that
we may have extracted only 50% of the DNA present in each sample. We are also aware that the
primers we chose almost certainly missed a proportion of the taxa present in soil. Assuming
only a 50% extraction of DNA and that we may have missed half of the taxa present in soil using
our primers, our estimates for the number of OTUs present in soil may be underestimated by a
factor of four. However, for the number of species proposed by Gans et al. (2005) to be correct,
our methods would have to have missed over 99.5% of the species in soil. Given the biases and
limitations of our methods described by others, this is simply not possible.
Comparison of species richness in each soil
The value of assessing diversity in a variety of soils has significant merit (Fig. 3). The
patterns of the rarefaction curves for the Canadian forest sample are very different from those of
the agricultural soils (Table 2). OTU abundance is particularly high at high levels of
dissimilarity while in the other samples the diversity is low at high levels of dissimilarity.
Conversely at low levels of dissimilarity the OTU abundance of the Canadian sample is lower
that that of the other samples. This result is interpreted as the forest sample being very phylum
rich but species poor while the samples are species rich and phylum poor. The cause of this
difference is unknown. Agricultural systems may reduce phylum diversity but increase
speciation. The Canadian sample may have higher phylum diversity because forest soils support
a wider diversity of bacteria. Speciation in this forext soil may be lower because cold
temperatures most of the year inhibits bacterial growth.
In contrast to the situation with bacterial diversity, the opposite is true regarding
crenarchaeota diversity. In Canada the number of sequences recovered was remarkably low and
highest in Illinois. Again, the cause of this is unknown.
The age of inexpensive, high throughput pyrosequencing allows the assessment of the
full taxonomic diversity of bacteria in soil for the first time. As a result, much better estimates of
OUT richness can be obtained from any sample. Here no attempt is made to define a species. A
significant debate exists in the literature regarding the species concept for bacteria. No judgment
is made on this here. However, it is clear that regardless of the % dissimilarity chosen by the
reader to define a particular level of taxonomic discrimination, the number of OTUs by any
definition cannot be as high as described by Gans et al. (2005).
REFERENCES
Borneman, J. and E.W. Triplett. 1997. Molecular microbial diversity in soils from eastern
Amazonia: evidence for unusual microorganisms and population shifts associated with
deforestation. Appl. Environ. Microbiol. 63:2647-2653.
Borneman, J., P.W. Skroch, J.A. Jansen, K.L. O'Sullivan, J.A. Palus, N.G. Rumjanek, J. Nienhuis,
and E.W. Triplett. 1996. Microbial diversity of an agricultural soil in Wisconsin. Appl. Environ.
Microbiol. 62:1935-1943.
DeSantis, T.Z., P. Hugenholtz, N. Larsen, M. Rojas, E.L. Brodie, K. Keller, T. Huber, D. Dalevi, P.
Hu, and G.L. Anderson. 2006. Greengenes, a chimera-checked 16S rRNA database and
workbench compatible with ARB. Appl. Environ. Microbiol. 72:5069-5072.
Edgar, R.C. 2004. MUSCLE: multiple sequence alignment with high accuracy and high
throughput. Nucleic Acids Research 32:1792-1797.
Felsenstein, J. 2005. PHYLIP (Phylogeny Inference Package) version 3.6. Distributed by the
author. Department of Genome Sciences, University of Washington, Seattle.
Felsenstein, J. 1989. PHYLIP - Phylogeny Inference Package (Version 3.2). Cladistics 5: 164166.
Fierer, N. and R.B. Jackson. 2006. The diversity and biogeography of soil bacterial
communities. Proc. Natl. Acad. Sci. (USA) 103;626-631.
Gans, J., M. Woilinsky, and J. Dunbar. 2005. Computational improvements reveal great
bacterial diversity and high metal toxicity in soil. Science 309:1387-1390.
Hong, S.-H., J. Bunge, S.-O. Jeon, and S.S. Epstein. Predicting microbial species richness.
Proc. Natl. Acad. Sci. (USA) 103:117-122.
Kent, A.D. and E.W. Triplett. 2002. Microbial communities and their interactions in soil and
rhizosphere ecosystems. Annu. Rev. Microbiol. 56:211-236.
Lane, D.J. 1991. 16S/23S rRNA sequencing. In: Stackebrandt E, Goodfellow M (eds) Nucleic
Acid Techniques in Bacterial Systematics. John Wiley & Sons, New York, pp 115–175.
Leininger, S., T. Urich, M. Schloter, L. Schwark, J. Qi, G.W. Nicol, J.I. Prosser, S.C. Schuster,
and C. Schleper. 2006. Archaea predominate among ammonia-oxidizing prokaryotes in soils.
Nature 442:806-809.
Margulies, M., M. Egholm, W.E. Altman, S. Attiya, J.S. Bader, L.A. Bemben, J. Berka, M.S.
Braverman, Y.-J. Chen, Z. Chen, S.B. Dewell, L. Du, J.M. Fierro, X.V. Gomes, B.C. Godwin,
W. He, S. Helgesen, C.H. Ho, G.P. Irzyk, S.C. Jando, M.L.I. Alenquer, T.P. Jarvie, K.B. Jirage,
J.-B. Kim, J.R. Knight, J.R. Lanza, J.H. Leamon, S.M. Lefkowitz, M. Lei, J. Li, K.L. Lohman,
H. Lu, V.B. Makhijani, K.E. McDade, M.P. McKenna, E.W. Myers, E. Nickerson, J.R. Nobile,
R. Plant, B.P. Puc, M.T. Ronan, G.T. Roth, G.J. Sarkis, J.F. Simons, J.W. Simpson, M.
Srinivasan, K.R. Tartaro, A.Tomasz, K.A. Vogt, G.A. Volkmer, S.H. Wang, Y. Wang, M.P.
Weiner, P. Yu, R.F. Begley, and J.M. Rothberg. 2005. Genome sequencing in microfabricated
high-density picolitre reactors. Nature 437:376-380
Moreira, D. 1998. Efficient removal of PCR inhibitors using agarose-embedded DNA
preparations. Nucleic Acids Res. 26,:3309-3310.
Mullins, T.D., T.B. Britschgi, R.L. Krest, and S.J. Giovannoni. 1995. Genetic comparisons
reveal the same unknown bacterial lineages in Atlantic and Pacific bacterioplankton
communities. Limnol. Oceanogr. 40:148–158.
Neufeld, J.D. and W.W. Mohn. 2005. Unexpectedly high bacteria diversity in arctic tundra
relative to boreal forest soils, revealed by serial analysis of ribosomal sequence tags. Appl.
Environ. Microbiol. 71:5710-5718.
Rupert, D., N. Cressie, and R.J. Carroll. 1989. A transformation/weighting model for estimating
Michaelis-Menton parameters. Biometrics 45:637-656.
Schloss, P.D. and J. Handelsman.
Computational Biology 2:786-793.
2006.
Toward a census of bacteria in soil.
PLOS
Schloss, P.D. and J. Handelsman. 2005. Introducing DOTUR, a computer program for defining
operational taxonomic units and estimating species richness. Appl. Environ. Microbiol.
71:1501-1506.
Sogin, M.L., H.G. Morrison, J.A. Huber, D.M. Welch, S.M. Huse, P.R. Neal, J.M. Arrieta, and
G.J. Herndl. 2006. Microbial diversity in the deep sea and the underexplored “rare biosphere”.
Proc. Natl. Acad. Sci. (USA) 103:12115-12120.
Torsvik, V., Goksoyr, and F.L. Daae. 1990. High diversity in DNA of soil bacteria. Appl.
Environ. Microbiol. 56:782-787.
Treusch, A.H., S. Leininger, A. Kletzin, S.C. Schuster, H.-P. Klenk and C. Schleper. 2005.
Novel genes for nitrite reductase and Amo-related proteins indicate a role of uncultivated
mesophilic crenarchaeota in nitrogen cycling. Environ. Microbiol. 7:1985-1995.
Tringe, S.G., C. von Mering, A. Kobayashi, A.A. Salamov, K. Chen, H. Chang, M. Podar, J.
Short, E. Mathur, J. Detter, et al. 2005. Comparative metagenomics of microbial communities.
Science 308:554-557.
Table 1. Locations, pH, locations, elevations, and types of soils as well as statistics on the 454
pyrosequencing.
Soil
Brazil
pH
4.6
lattitue/longitude -29.539671S, 55.107556W
Florida
7.0
26.663199N,
-80.628500W
Illinois
6.4
40.104616N,
-88.226517W
Canada
5.4
52.743203N,
- 91.718433W
elevation (m)
soil
classification
132
Distrophic oxisol
224
mesic Aquic
Argiudoll
314
Dystric Brunisol
GenBank
accession
numbers
No. of bacterial
16S rRNA gene
sequences
No.
of
crenarchaeotal
16S rRNA gene
sequences
Maximum read
length (bps)
Average
read
length (bps)
EF222481EF248596
2
Lauderhill euic
hyperthermic
Lithic
EF248597EF276844
EF276845EF308590
EF308591EF361836
26,116
28,248
31,746
53,246
1,259
3,546
4,530
5
141
159
144
160
103
102
104
102
Table 2. Estimated number of OTUs for each sample using parametric (rarefaction) and nonparametric estimators (Chao1 and Ace). The number of sequences required to reach 90% and
95% of the maximum OTU level at each % dissimilarity was calculated using the rarefaction
estimator.
Brazil
Florida
Illinois
Canada
%
dissimilarity
Rarefaction
Chao1
Ace
0
3
5
10
20
0
3
5
10
20
0
3
5
10
20
0
3
5
10
20
12,196
3,611
2,395
648
298
16,485
4,477
3,033
1,061
529
15,797
4.010
2,651
926
538
20,920
15,188
13,805
10,604
9,376
25,972
5,021
3,183
756
338
34,380
5,666
3,686
1,110
509
43,652
6,040
3,403
1,057
537
20,244
20,244
12,282
10,230
9,246
26,395
4,888
3,046
767
344
37,537
5,820
3,753
1,198
539
51,378
5,890
3,553
1,154
575
20,596
13,329
12,234
10,213
9,236
Sequences
needed for
90% max.
OTUs
209,864
113,205
98,495
68,172
65,500
266,429
127,715
104,457
85,361
82,610
583,003
282,214
247,373
228,840
263,225
842,692
762,838
735,323
661,496
617,608
Sequences
needed for
95% max.
OTUs
443,047
238,988
207,934
143,919
138,279
562,462
269,621
220,514
180,207
174,398
615,392
297,892
261,116
241,554
277,849
1,779,017
1,610,438
1,552,349
1,396,492
1,303,840
Figure 1. Google earth images of the four locations used for soil collection.
Fig. 2. The effect of the number of sequences available on the estimation of the number of
OTUs.
Fig. 3. Rarefaction curves depicting the effect of % dissimilarity on the number of OTUs
identified. Note the comparatively high species richness of the agricultural samples while the
Canadian forest soil has very high phylum richness.
FIGURE 4. Relative abundance of phylogenetic divisions for each soil library in which 16S
sequences were classified according to the nearest neighbor in the Greengenes database
(http://greengenes.lbl.gov).