Assessing Optimum Soil Sample Size for Fungal and Bacterial Community Structure Analyses Using High Throughput Sequencing of rRNA Gene Amplicons

We examined the effect of different soil sample sizes obtained from an agricultural field on bacterial and fungal community structure indices. DNA extracted from soil sample sizes of 0.25, 1, 5 and 10 g using MoBIO kits and from 10 and 100 g sizes using a proprietary method (SARDI) were used as templates for high-throughput sequencing of 16S and 28S rRNA gene amplicons for bacteria and fungi, respectively, on the on the Illumina MiSeq and Roche 454 platforms. Sample size significantly affected overall bacterial and fungal community structure, replicate dispersion and the number of operational taxonomic units (OTUs) retrieved. Richness, evenness and diversity were also significantly affected. The largest diversity estimates were always associated with the 10 g MoBIO extractions with a corresponding reduction in replicate dispersion. There were no consistent biases towards any bacterial taxa, but bias was evident in the fungal data. Specifically, smaller MoBIO extractions were biased towards many unclassified Eukaryota incertae sedis and unclassified glomeromycota while the SARDI method retrieved more abundant OTUs containing unclassified Pleosporales and the fungal genera Alternaria and Cercophora. Overall, these findings indicate that a 10 g soil DNA extraction is most suitable for both soil bacterial and fungal communities for retrieving optimal diversity while decreasing replicate variation.