Fine-scale bacterial beta diversity within a complex ecosystem (Zodletone Spring, OK, USA): the role of the rare biosphere - PubMed (original) (raw)
Fine-scale bacterial beta diversity within a complex ecosystem (Zodletone Spring, OK, USA): the role of the rare biosphere
Noha H Youssef et al. PLoS One. 2010.
Abstract
Background: The adaptation of pyrosequencing technologies for use in culture-independent diversity surveys allowed for deeper sampling of ecosystems of interest. One extremely well suited area of interest for pyrosequencing-based diversity surveys that has received surprisingly little attention so far, is examining fine scale (e.g. micrometer to millimeter) beta diversity in complex microbial ecosystems.
Methodology/principal findings: We examined the patterns of fine scale Beta diversity in four adjacent sediment samples (1mm apart) from the source of an anaerobic sulfide and sulfur rich spring (Zodletone spring) in southwestern Oklahoma, USA. Using pyrosequencing, a total of 292,130 16S rRNA gene sequences were obtained. The beta diversity patterns within the four datasets were examined using various qualitative and quantitative similarity indices. Low levels of Beta diversity (high similarity indices) were observed between the four samples at the phylum-level. However, at a putative species (OTU(0.03)) level, higher levels of beta diversity (lower similarity indices) were observed. Further examination of beta diversity patterns within dominant and rare members of the community indicated that at the putative species level, beta diversity is much higher within rare members of the community. Finally, sub-classification of rare members of Zodletone spring community based on patterns of novelty and uniqueness, and further examination of fine scale beta diversity of each of these subgroups indicated that members of the community that are unique, but non novel showed the highest beta diversity within these subgroups of the rare biosphere.
Conclusions/significance: The results demonstrate the occurrence of high inter-sample diversity within seemingly identical samples from a complex habitat. We reason that such unexpected diversity should be taken into consideration when exploring gamma diversity of various ecosystems, as well as planning for sequencing-intensive metagenomic surveys of highly complex ecosystems.
Conflict of interest statement
Competing Interests: The authors have declared that no competing interests exist.
Figures
Figure 1. Distribution of phyla in Zodletone spring source sediment for the 4 quadrants studied.
Percentage abundance is shown on the Y-axis for (A) abundant phyla with >1% abundance in each quadrant studied, (B) phyla with % abundance ranging between 0.1 and 1% in quadrant 1, (C) phyla with % abundance ranging between 0.01 and 0.1% in quadrant 1, and (D) phyla with % abundance <0.01% in quadrant 1. Color-coding is as follows: quadrant 1 (blue), quadrant 2 (red), quadrant 3 (green), and quadrant 4 (purple).
Figure 2. Venn diagrams showing the number of unique and shared OTUs between the 4 quadrants studied for (A) empirical cutoff (n = 1) for defining rare species, (B) empirical cutoff (n≤2) for defining rare species, (C) empirical cutoff (n≤5) for defining rare species, (D) empirical cutoff (n≤10) for defining rare species, (E) empirical cutoff (n>10) for defining abundant species, (F) empirical cutoff (% abundance ≤0.004%) for defining rare species, (G) empirical cutoff (% abundance >1%) for defining abundant species, and (H) the entire datasets.
Color-coding is as follows: quadrant 1 (yellow), quadrant 2 (red), quadrant 3 (green), and quadrant 4 (blue).
Figure 3. Non-metric multidimensional scaling plots of the 4 quadrants studied.
Pair wise Bray-Curtis similarity indices for each of the quadrants at 4 empirical rare cutoffs (n = 1 (N1), n≤2 (N2), n≤5 (N5), and n≤10 (N10)), one empirical abundant cutoff (n>10 (Ab)), as well as for the entire datasets (All) were used to construct the plot. Grey shapes show rare and abundant members from each quadrant clustering closer together with the whole quadrant. Stress value was close to 0, and most of the variation (65–95.3%) was explained by 2 axes. The addition of a third axis did not improve the model substantially.
Figure 4. Effect of inadequate sampling on the observed beta diversity in Zodletone spring.
A sub-sampling approach was implemented in which 3 subsamples from quadrant 2 (sub1, sub2, and sub3) were compared to three subsamples, each from a different quadrant (Quad 1_r, Quad 3_r, and Quad 4_r). (A) The percentage unshared OTUs, and the percentage shared OTUs between 2 quadrants, and 3 quadrants for the 3 subsamples from quadrant 2 versus the 3 subsamples from quadrants 1, 3, and 4 (* denotes that the percentage shared/unshared OTUs were significantly different between the 2 groups). Error bars represent standard deviations from 3 pairs of subsamples (2 quads) and 3 subsamples (1 quad). No error bars were obtained for the single 3-quads data point. Color-coding is as follows: the 3 subsamples from quadrant 2 (black), and the 3 subsamples from quadrants 1, 3, and 4 (white). (B) Non-metric multidimensional scaling plot for the 6 subsamples studied. Pair wise Bray-Curtis similarity indices were used to construct the plot. Stress value was close to 0 and most of the variation (77.8%) was explained by 2 axes. (C) Observed and expected number of species (Sobs, and Sexpected, respectively) and singletons for each of the quadrants studied. The numbers were used to construct theoretical Venn diagrams as described in text. (D) The percentage shared OTUs between 4 quadrants, 3 quadrants, 2 quadrants, and the percentage unshared OTUs calculated from the obtained Venn diagram for n = 1 (shown in Figure 2A) (observed data), and the average percentages calculated from the theoretical Venn diagrams (predicted data). Error bars represent standard deviations from 4 data points for each of the observed and predicted unshared OTUs percentages (1 quad), 6 data point for the observed shared OTUs percentages and 4 data points for the predicted shared OTUs percentages between 2 quads, 4 data points for each of the observed and predicted shared OTUs percentages between 3 quads, 4 data points for the predicted shared OTUs percentages between 4 quads. No error bars were obtained for the single 4-quads observed shared OTUs percentage data point. Color-coding is as follows: Observed data (black), and predicted data (white).
References
- Acosta-Martínez V, Dowd S, Sun Y, Allen V. Tag-encoded pyrosequencing analysis of bacterial diversity in a single soil type as affected by management and land use. Soil Biol Biochem. 2008;40:2762–2770.
- Chariton AA, Court LN, Hartley DM, Colloff MJ, Hardy CM. Ecological assessment of estuarine sediments by pyrosequencing eukaryotic ribosomal DNA. Front Ecol Env. 2010;8:233–238.
Publication types
MeSH terms
Substances
LinkOut - more resources
Full Text Sources
Other Literature Sources