Random Sampling Process Leads to Overestimation of β-Diversity of Microbial Communities

Abstract

The site-to-site variability in species composition, known as β-diversity, is crucial to understanding spatiotemporal patterns of species diversity and the mechanisms controlling community composition and structure. However, quantifying β-diversity in microbial ecology using sequencing-based technologies is a great challenge because of a high number of sequencing errors, bias, and poor reproducibility and quantification. Herein, based on general sampling theory, a mathematical framework is first developed for simulating the effects of random sampling processes on quantifying β-diversity when the community size is known or unknown. Also, using an analogous ball example under Poisson sampling with limited sampling efforts, the developed mathematical framework can exactly predict the low reproducibility among technically replicate samples from the same community of a certain species abundance distribution, which provides explicit evidences of random sampling processes as the main factor causing high percentages of technical variations. In addition, the predicted values under Poisson random sampling were highly consistent with the observed low percentages of operational taxonomic unit (OTU) overlap (<30% and <20% for two and three tags, respectively, based on both Jaccard and Bray-Curtis dissimilarity indexes), further supporting the hypothesis that the poor reproducibility among technical replicates is due to the artifacts associated with random sampling processes. Finally, a mathematical framework was developed for predicting sampling efforts to achieve a desired overlap among replicate samples. Our modeling simulations predict that several orders of magnitude more sequencing efforts are needed to achieve desired high technical reproducibility. These results suggest that great caution needs to be taken in quantifying and interpreting β-diversity for microbial community analysis using next-generation sequencing technologies.

Early detection of invasive species in marine ecosystems using high-throughput sequencing: technical challenges and possible solutions

“…1). Despite the fact that PCR-free strategies such as DNA capture and direct shotgun sequencing (i.e., metagenomics) have been proposed to avoid biased amplification and/ or due to the lack of universal PCR primers (Taberlet et al 2012), these methods have not been well validated and/ or have limited capacities to recover rare taxa (Zhou et al 2013a;Srivathsan et al 2015;. We therefore focus on the commonly used PCR-method strategy (i.e., metabarcoding strategy; Fig.…”

Temperature, inocula and substrate: Contrasting electroactive consortia, diversity and performance in microbial fuel cells

“…However, these techniques still only offer a partial view of diversity in a microbial system, and the diversity observed is still typically a function of the number of sequences analysed [32]. These difficulties can be overcome using Bayesian techniques that infer the underlying species abundance curves which in turn can be used to estimate the total diversity of the sample [33,34].…”

Rare biosphere exploration using high-throughput sequencing: research progress and perspectives

“…When reproducibility was examined based on the abundance of OTUs, the reproducibility of low-abundance OTUs was much lower than high-abundance ones (Panel II), for example, as low as \25 % for singletons (OTUs represented by one sequence) versus 100 % for OTUs with the number of sequences [100 (Zhan et al 2014c). Both laboratory work and a newly developed mathematical framework support the view that low reproducibility was a result of random sampling process during both biological sample preparation and data generation Ihrmark et al 2012;Zhou et al 2013b;Zhan et al 2014c). For biological sample collection and preparation, low population density may lead to inconsistent presence of rare species in subsamples.…”

“…Random sampling of rare taxa: a major cause for low reproducibility (Type I error) Recent tests showed that the reproducibility of metabarcoding data was low among parallel replicates, for example, \30 % of OTUs for soil microbes when using two or three tagged replicates (Zhou et al 2013b). When reproducibility was examined based on the abundance of OTUs, the reproducibility of low-abundance OTUs was much lower than high-abundance ones (Panel II), for example, as low as \25 % for singletons (OTUs represented by one sequence) versus 100 % for OTUs with the number of sequences [100 (Zhan et al 2014c).…”