Microbial Communities as Experimental Units

Abstract

Artificial ecosystem selection is an experimental technique that treats microbial communities as though they were discrete units by applying selection on community-level properties. Highly diverse microbial communities associated with humans and other organisms can have significant impacts on the health of the host. It is difficult to find correlations between microbial community composition and community-associated diseases, in part because it may be impossible to define a universal and robust species concept for microbes. Microbial communities are composed of potentially thousands of unique populations that evolved in intimate contact, so it is appropriate in many situations to view the community as the unit of analysis. This perspective is supported by recent discoveries using metagenomics and pangenomics. Artificial ecosystem selection experiments can be costly, but they bring the logical rigor of biological model systems to the emerging field of microbial community analysis.

Figure 1

Group selection experiments. (a) Choose units of selection with a variable phenotype (ϕ) from a common stock. Maintain parent populations (P) under identical environmental conditions. Select a fraction of each of the parent populations that express a phenotypic value at the upper end of the frequency distribution (for high line). The null line is a control that excludes the possibility of group selection. Random line units are selected randomly (Sr ϕ) as an additional control. We would usually predict no significant change in ϕ̄ the null or random lines. Mix the selected microcosms to produce new populations (O₁). Repeat using the offspring populations as the new parents (O_n). (b) The response to selection is a statistically significant difference between the mean phenotype (ϕ̄) of the high and low lines, shown here as nonoverlapping confidence intervals. (c) The main distinction between individual and group selection experiments is the unit of selection, the definition of ϕ, and the calculation of ϕ̄. Groups of organisms form the unit of selection and ϕ is a property of individual groups. ϕ̄ is a property of a group of groups.

Figure 2

Microbial communities as lineages. Reciprocal transplant of the gut communities of mice reproduces the function of the transplanted community in the new host (summary of results from Vijay-Kumar et al. 2010). Sterile embryos of mutant mice received their gut microbial communities during birth by wild-type mothers. Wild-type mice raised under sterile conditions received their gut community through feeding by the experimenters. Abbreviations: MS+, mutant mice lacking an active Toll-like receptor 5 and exhibiting metabolic syndrome; MSp, partial manifestation of metabolic syndrome; WT, wild-type mice without metabolic syndrome.

Figure 3

Low dilution rate preserves diversity. Mathematical models explain why enrichment is not the inevitable outcome of periodic or continuous dilution in microcosms. If the dilution rate is low enough or the mutation rate high enough, multiple genotypes can persist. This plot shows the steady-state solution of a system of differential equations across a range of values for mutation rate and dilution rate. Source: Figure redrawn with permission from model results reported by Gudelj and colleagues (2007).

Figure 4

Artificial ecosystem selection experiments share some similarities with other more established microbiology techniques. Enrichment, direct plating, and artificial ecosystem selection all perform dilution of a diverse community, but there are crucial differences among them. The most critical differences are the rate of dilution events and the composition of the growth medium. Another technique commonly used to study diverse communities is long-term treatment. Treatment experiments can be performed in laboratory microcosms or in natural environments. An environmental parameter such as addition of fertilizers is varied across a replicate-block experimental design and the resident communities are allowed to adjust for a relatively long period (e.g., Chu et al. 2007). In in situ soil treatment experiments, the dilution rate is necessarily zero.