Sampling forest soils to describe fungal diversity and composition. Which is the optimal sampling size in mediterranean pure and mixed pine oak forests?

Abstract

Soil sampling is a critical step affecting perceived fungal diversity, however sampling optimization for high-throughput-DNA sequencing studies have never been tested in Mediterranean forest ecosystems. We identified the minimum number of pooled samples needed to obtain a reliable description of fungal communities in terms of diversity and composition in three different Mediterranean forests (pine, oak, and mixed-pine-oak). Twenty soil samples were randomly selected in each of the three plots per type. Samples obtained in 100 m2 plots were pooled to obtain mixtures of 3, 6, 10, 15, 20 samples, and sequenced using Illumina MiSeq of fungal ITS2 amplicons. Pooling three soil samples in Pinus and Quercus stands provided consistent richness estimations, while at least six samples were needed in mixed-stands. β-diversity decreased with increasing sample pools in monospecific-stands, while there was no effect of sample pool size on mixed-stands. Soil sample pooling had no effect over species composition. We estimate that three samples would be already optimal to describe fungal richness and composition in Mediterranean pure stands, while at least six samples would be needed in mixed stands.