Quantitative Comparisons of 16S rRNA Gene Sequence Libraries from Environmental Samples

Abstract

To determine the significance of differences between clonal libraries of environmental rRNA gene sequences, differences between
homologous coverage curves, CX(D), and heterologous coverage curves, CXY(D), were calculated by a Cramér-von Mises-type statistic and compared by a Monte Carlo test procedure. This method successfully
distinguished rRNA gene sequence libraries from soil and bioreactors and correctly failed to find differences between libraries
of the same composition.

Full text

APPLIED AND ENVIRONMENTAL MICROBIOLOGY,
0099-2240/01/$04.00⫹0 DOI: 10.1128/AEM.67.9.4374–4376.2001
Sept. 2001, p. 4374–4376 Vol. 67, No. 9
Copyright © 2001, American Society for Microbiology. All Rights Reserved.
Quantitative Comparisons of 16S rRNA Gene Sequence
Libraries from Environmental Samples
DAVID R. SINGLETON,
1
MICHELLE A. FURLONG,
1
STEPHEN L. RATHBUN,
2
AND WILLIAM B. WHITMAN
1
*
Departments of Microbiology
1
and Statistics,
2
University of Georgia,
Athens, Georgia 30602-2605
Received 8 March 2001/Accepted 11 June 2001
To determine the significance of differences between clonal libraries of environmental rRNA gene sequences,
differences between homologous coverage curves, C
X
(D), and heterologous coverage curves, C
XY
(D), were cal-
culated by a Crame´r-von Mises-type statistic and compared by a Monte Carlo test procedure. This method
successfully distinguished rRNA gene sequence libraries from soil and bioreactors and correctly failed to find
differences between libraries of the same composition.
The sequencing of 16S rRNA genes from clone libraries of
DNAs from environmental samples has led to a wealth of
information concerning prokaryotic diversity. However, in ad-
dition to methodological problems in producing libraries rep-
resentative of the environmental sample (for a review, see
reference 8), this approach is also limited by the difficulty in
comparing libraries and determining if they are significantly
different.
This problem can be addressed quantitatively by application
of the formula for coverage as described by Good (4). Let Xbe
a collection of sequences, such as a library of 16S rRNA genes.
Define the “homologous” coverage of X(or C
X
) by a sample
from Xto be C
X
⫽1⫺(N
X
/n), where N
X
is the number of
unique sequences in the sample (i.e., sequences without a
replicate) and nis the total number of sequences. In practice,
the definition of N
X
depends upon the criteria used to define
uniqueness. For instance, McCaig et al. (6) considered se-
quences without a homolog of ⱖ97% similarity to be unique.
Other authors have used ⱖ99% sequence similarity as the
criterion. In principle, uniqueness can be defined at any level
of sequence similarity or evolutionary distance (D) and a “ho-
mologous coverage curve,” or C
X
(D), can be generated by
plotting C
X
versus D(Fig. 1). The coverage curve then de-
scribes how well the sample represents the entire library Xat
various levels of relatedness. Typically, coverage might be low
at high levels of relatedness (low values of D), indicating that
only a small fraction of the sequences representing unique
species are, in fact, sampled. In contrast, coverage might be
much higher at low levels of relatedness, indicating that rep-
resentatives of most of the deep phylogenetic groups present in
Xare found in the sample.
While C
X
is the “homologous coverage” of Xby a sample of
X, it is also possible to calculate a “heterologous coverage” of
X(or C
XY
) by a sample Yfrom another collection of sequences
by the following formula: C
XY
⫽1⫺(N
XY
/n), where N
XY
is the
number of sequences in a sample of Xthat are not found in a
sample of Yand nis the number of sequences in the sample of
X. Similarly to N
X
,N
XY
can also be defined at different levels
of Dto generate a coverage curve, C
XY
(D). Moreover, if X⫽
Y, one might expect the coverage curves C
X
(D) and C
XY
(D) [as
well as C
Y
(D) and C
YX
(D)] to be similar. Thus, a test for dif-
ferences between these coverage curves is also a test for dif-
ferences between Xand Y. To determine if the coverage curves
C
X
(D) and C
XY
(D) are significantly different, the distance be-
tween the two curves are first calculated by using the Crame´r-
von Mises test statistic (7):
⌬CXY ⫽
冘
D⫽0.0
0.5
共CX⫺CXY兲2
where Dincreases in increments of 0.01. If X⫽Y, then ⌬C
XY
should not be significantly different than a ⌬Ccalculated after
randomly shuffling sequences between the two samples, Xand
Y. Typically, the sequences are randomly shuffled a large num-
ber (N) of times (e.g., N⫽999) and ⌬C
XY
is calculated after
each shuffling. The randomized values plus the empirical value
of ⌬C
XY
are ranked from largest to smallest, and then the P
value is estimated to be r/(N⫹1), where rdenotes the rank of
the empirical value of ⌬C
XY
(5). The two libraries are consid-
ered significantly different when P⬍0.05. We have created a
computer program (LIBSHUFF) that uses a sorted distance
matrix containing both Xand Yas input and returns the cov-
erage curves C
X
(D), C
Y
(D), C
XY
(D), and C
YX
(D), as well as the
Pvalues for both ⌬C
XY
and ⌬C
YX
, from the distribution of ⌬C.
In addition, the distribution of (C
X
⫺C
XY
)
2
with Dappears to
be informative and is given as well (see below). The computer
program LIBSHUFF was written in Perl and can be down-
loaded along with more detailed instructions on its use at http:
//www.arches.uga.edu/⬃whitman/libshuff.html.
A first test of this method was done to ensure that samples
from the same library were not shown to be different. Thus, a
collection of clonal sequences (n⫽275) from a soil community
study (6) was divided into two samples based upon accession
numbers (138 odds and 137 evens). Although the study con-
tained sequences from two sample sites (SL and SAF clones),
sequences from both sites were placed in each data set to form
nearly equivalent samples. A comparison of ⌬C
odds/evens
to ⌬C
* Corresponding author. Mailing address: Department of Microbi-
ology, University of Georgia, 527 Biological Sciences Bldg.; Athens,
GA 30602-2605. Phone: (706) 542-4219. Fax: (706) 542-2674. E-mail:
whitman@arches.uga.edu.
4374

values resulted in P⫽0.871, which indicated that the two
samples were not significantly different (Fig. 1A). Similar re-
sults were obtained for ⌬C
evens/odds
and other arbitrarily di-
vided sequence libraries (Table 1). Thus, as expected, samples
taken from the same library were not found to be different.
To demonstrate that this procedure could correctly differ-
entiate samples from different libraries, sequences of clones
obtained from an activated sludge (SBR1; n⫽97; reference 1)
were compared to grassland soil SL clones. The SBR1 clones
were found to be significantly different from the SL clones
(P⫽0.001; Fig. 1B). More information on the nature of this
difference was obtained by examination of the distribution of
(C
X
⫺C
XY
)
2
with D(Fig. 1B). At low D, the actual (C
X
⫺
C
XY
)
2
exceeded the comparable values at P⫽0.05 obtained
during the calculation of ⌬C. This result suggested that the
libraries differed greatly at D⬍0.10 but shared many deep
taxa. However, smaller differences at D⬎0.3 suggested that
not all deep phylogenetic groups were found in both libraries.
Similar results were also obtained for comparisons of other soil
and bioreactor libraries (Table 1 and data not shown).
Three sequence collections consisting of multiple samples
were analyzed to determine if differences between the samples
could be detected (Table 1). Clonal libraries derived from the
microbial populations of phosphate-removing (SBR1) and
non-phosphate-removing (SBR2) bioreactors differed in the
abundance of certain taxa (1). However, these differences were
not shown to be significant by our method (Table 1). The
compositions of libraries from the microbial communities of
improved (SL) and unimproved (SAF) upland grass pasture
soils were not found to be significantly different (6). We also
obtained the same conclusion by our method (Table 1). Fi-
nally, comparisons of restriction fragment length types from
C0 and S0, two clonal libraries derived from arid soils, sug-
gested that C0 was more diverse than S0 (2). Our analysis of
the sequences obtained from this study was consistent with this
conclusion and further suggested that S0 was a subset of C0.
⌬C
S0/C0
was not significant, which suggested that all of the taxa
present in S0 were also present in C0 (Table 1). However, the
reciprocal value ⌬C
C0/S0
was significant; therefore, C0 also
contained sequences of one or more taxa not found in S0. The
distribution of (C
X
⫺C
XY
)
2
with Dfurther indicated that the
additional taxa in C0 represented moderately deep phyloge-
netic groups, 0.15 ⬍D⬍0.25 (Fig. 1C).
FIG. 1. Results of selected LIBSHUFF comparisons. Homologous
(E) and heterologous (F) coverage curves for 16S rRNA gene se-
quence libraries from environmental samples are shown. Solid lines
indicate the value of (C
X
⫺C
XY
)
2
for the original samples at each value
of D.Dis equal the Jukes-Cantor evolutionary distance determined
by the DNADIST program of PHYLIP (3). Broken lines indicate the
950th value (or P⫽0.05) of (C
X
⫺C
XY
)
2
for the randomized samples.
(A) Comparison of clones from grassland soils with odd (X) and even
(Y) accession numbers. (B) Comparison of bioreactor clones SBR1 (X)
and grassland soil SL clones (Y). (C) Comparison of C0 (X) and S0 (Y)
clones from arid soils.
TABLE 1. Comparisons of environmental clone libraries
Site (reference)
Homologous
(X)Heterologous
(Y)P
b
Clones nClones
Grassland soils (6) Odds
a
138 Evens
a
0.871
Evens
a
137 Odds
a
0.933
SAF 138 SL 0.120
SL 137 SAF 0.135
Bioreactors (1) Odds
a
95 Evens
a
0.853
Evens
a
94 Odds
a
0.623
SBR1 97 SBR2 0.308
SBR2 92 SBR1 0.824
Arid soils (2) Odds
a,c
56 Evens
a
0.251
Evens
a
56 Odds
a,c
0.516
C0 59 S0 0.042
S0 53 C0 0.398
Grassland soil/bioreactor SAF 138 SBR1 0.001
SBR1 97 SAF 0.002
SL 137 SBR1 0.001
SBR1 97 SL 0.001
a
Sequences with odd or even accession numbers. Contains mixtures of both
libraries described in the reference, and they are not expected to be different.
b
Value of r/(N⫹1) as described in the text.
c
Accession number AF128647 could not be found and was not included.
VOL. 67, 2001 QUANTITATIVE COMPARISONS OF SEQUENCE LIBRARIES 4375

Sample size should have a major effect on comparisons of
libraries. The minimum number of sequences necessary to dis-
tinguish two dissimilar libraries was expected to increase with
the complexity of the libraries and decrease with the magni-
tude of the dissimilarity. This point was examined in detail by
using two libraries of high diversity and dissimilarity. Variable
numbers of clonal sequences were randomly selected from
either library SBR1 or SL (Y) and compared to the opposite
library (X), and Pvalues were determined for 10 replicates.
Approximately 20 and 25 sequences from SBR1 and SL, re-
spectively, were required to differentiate the two libraries (P⬍
0.05) when Xwas represented by 97 and 137 sequences, re-
spectively (Fig. 2). Tests were also performed to investigate the
required sample size of X(SBR1) when the size of Y(SL) was
small. It was found that nearly all (ⱖ90) of the sequences from
the SBR1 library were required to distinguish these libraries
when the SL library (Y) was represented by 20 sequences (data
not shown). When the sizes of both libraries were varied, they
were consistently detected as different when the SBR1 (X) and
SL (Y) libraries were represented by ⱖ40 and ⱖ30 sequences,
respectively (data not shown). While these results may not
generalize to all environmental samples, they should be repre-
sentative of comparisons of libraries from diverse communi-
ties, such as those found in soil and bioreactors. Importantly,
these results suggest than modestly sized libraries from micro-
bial communities similar in complexity to those used in this
study will be distinguished by this method.
We thank Kamyar Farahi and Rob Waldo for help with program-
ming in Perl. We also thank Lihua Wang of the Statistical Consulting
Office at the University of Georgia for help.
This work was supported in part by an award from the Division of
Molecular and Cellular Biosciences at NSF (MCB-0084164).
REFERENCES
1. Bond, P. L., P. Hugenholtz, J. Keller, and L. L. Blackall. 1995. Bacterial
community structures of phosphate-removing and non-phosphate-removing
activated sludges from sequencing batch reactors. Appl. Environ. Microbiol.
61:1910–1916.
2. Dunbar, J., S. Takala, S. M. Barns, J. A. Davis, and C. R. Kuske. 1999. Levels
of bacterial community diversity in four arid soils compared by cultivation and
16S rRNA gene cloning. Appl. Environ. Microbiol. 65:1662–1669.
3. Felsenstein, J. 1993. PHYLIP (phylogenetic inference package) version 3.5c.
University of Washington, Seattle.
4. Good, I. J. 1953. The population frequencies of species and the estimation of
population parameters. Biometrika 40:237–264.
5. Hope, A. C. A. 1968. A simplified Monte Carlo significance test procedure.
J. Royal Statist. Soc. B 30:582–598.
6. McCaig, A. E., L. A. Glover, and J. I. Prosser. 1999. Molecular analysis of
bacterial community structure and diversity in unimproved and improved
upland grass pastures. Appl. Environ. Microbiol. 65:1721–1730.
7. Pettitt, A. N. 1982. Cramer-von Mises statistic, p. 220–221. In S. Kotz and
N. L. Johnson (ed.), Encyclopedia of statistical sciences. Wiley-Interscience,
New York, N.Y.
8. von Wintzingerode, F., U. B. Go¨bel, and E. Stackebrandt. 1997. Determina-
tion of microbial diversity in environmental samples: pitfalls of PCR-based
rRNA analysis. FEMS Microbiol. Rev. 21:213–229.
FIG. 2. Effect of sample size on the discrimination of libraries. A
comparison of the SL library from grassland soil (Y;n⫽variable) to
the bioreactor library SBR1 (X;n⫽97) (F) and a comparison of the
SBR1 (Y;n⫽variable) library to the SL (X;n⫽137) library (E)
shown. Each point represents an average of 10 replicates, and the error
bars are 1 standard deviation. The broken line indicates P⫽0.05.
4376 SINGLETON ET AL. APPL.ENVIRON.MICROBIOL.