Two-stage microbial community experimental design

Abstract

Microbial community samples can be efficiently surveyed in high throughput by sequencing markers such as the 16S ribosomal RNA gene. Often, a collection of samples is then selected for subsequent metagenomic, metabolomic or other follow-up. Two-stage study design has long been used in ecology but has not yet been studied in-depth for high-throughput microbial community investigations. To avoid ad hoc sample selection, we developed and validated several purposive sample selection methods for two-stage studies (that is, biological criteria) targeting differing types of microbial communities. These methods select follow-up samples from large community surveys, with criteria including samples typical of the initially surveyed population, targeting specific microbial clades or rare species, maximizing diversity, representing extreme or deviant communities, or identifying communities distinct or discriminating among environment or host phenotypes. The accuracies of each sampling technique and their influences on the characteristics of the resulting selected microbial community were evaluated using both simulated and experimental data. Specifically, all criteria were able to identify samples whose properties were accurately retained in 318 paired 16S amplicon and whole-community metagenomic (follow-up) samples from the Human Microbiome Project. Some selection criteria resulted in follow-up samples that were strongly non-representative of the original survey population; diversity maximization particularly undersampled community configurations. Only selection of intentionally representative samples minimized differences in the selected sample set from the original microbial survey. An implementation is provided as the microPITA (Microbiomes: Picking Interesting Taxa for Analysis) software for two-stage study design of microbial communities.

Inflammatory Bowel Disease as a Model for Translating the Microbiome

“…Staged study designs are particularly appropriate for the microbiome (Tickle et al, 2013), as they allow the benefits of large cross-sectional profiles combined with molecular detail derived from a subset of individuals, conditions, or time points.…”

“…“Bottom up” molecular detail can be added efficiently by perturbation studies in cell lines, but these are limited in relatability to primary populations. Two-stage study designs offer cost-effective scalability to larger cohorts (Tickle et al, 2013), which are also better powered for genetic profiling. Patient samples can be used to derive primary cell lines (Miyoshi and Stappenbeck, 2013) or organoids (Sato et al, 2011) for controlled perturbations that more closely recapitulate in vivo conditions.…”

Experimental metagenomics and ribosomal profiling of the human skin microbiome

“…[18][19][20] For these reasons, it is crucial to have properly powered cohorts and ad hoc longitudinal study designs. [21][22][23][24] Propionibacterium acnes and S. epidermidis are two prevalent members of the skin microbiome, but many other bacteria from the Actinobacteria, Firmicutes, Bacteroidetes and Proteobacteria phyla are frequently found. [3,23] The micro-eukaryotic component of the microbiome also comprises many organisms (most notably Malassezia spp.…”

Sequencing and beyond: integrating molecular 'omics' for microbial community profiling

“…Combining the lower cost and decreased resolution of amplicon sequencing with the higher cost and increased resolution of WMS sequencing (here, one “unit” of WMS sequencing and four “units” of amplicon sequencing are considered to have equivalent costs) enables richer experimental designs. For example, two-stage study designs begin by surveying a large number of individuals using amplicon sequencing and then follow-up with a subset of samples using WMS sequencing (selected based on individuals that are representative of the group or those that represent the extreme cases within the group 119 ). Similarly, time course studies can combine amplicon sequencing, which is used to survey a large number of time points, with WMS sequencing, which is applied to analyze a subset of time points (such as the first and last) in greater detail.…”