Assessment of microbial α-diversity in one meter squared topsoil

Due to the tremendous diversity of microbial organisms in topsoil, the estimation of saturated richness in a belowground ecosystem is still challenging. Here, we intensively surveyed the 16S rRNA gene in four 1 m2 sampling quadrats in a typical grassland, with 141 biological or technical replicates generating over 11 million sequences per quadrat. Through these massive data sets and using both non-asymptotic extrapolation and non-parametric asymptotic approaches, results revealed that roughly 15 919±193, 27 193±1076 and 56 985±2347 prokaryotic species inhabited in 1 m2 topsoil, classifying by DADA2, UPARSE (97% cutoff) and Deblur, respectively, and suggested a huge difference among these clustering tools. Nearly 500 000 sequences were required to catch 50% species in 1 m2, while any estimator based on 500 000 sequences would still lose about a third of total richness. Insufficient sequencing depth will greatly underestimate both observed and estimated richness. At least ∼911 000, ∼3461 000, and ∼1 878 000 sequences were needed for DADA2, UPARSE, and Deblur, respectively, to catch 80% species in 1 m2 topsoil, and the numbers of sequences would be nearly twice to three times on this basis to cover 90% richness. In contrast, α-diversity indexes characterized by higher order of Hill numbers, including Shannon entropy and inverse Simpson index, reached saturation with fewer than 100 000 sequences, suggesting sequencing depth could be varied greatly when focusing on exploring different α-diversity characteristics of a microbial community. Our findings were fundamental for microbial studies that provided benchmarks for the extending surveys in large scales of terrestrial ecosystems.