Abstract: Soil fungi are known to be important members of ecosystems; they are mutualists and parasites of plants, decomposers of organic matter, and can have indirect effects on herbivores. There are thousands of soil fungal operational taxonomic units (OTUs) identified in a single soil sample, but despite their importance and prolificity, soil microorganisms are understudied due to the difficulty of obtaining species-level matches and generating ecological meaning from complex sequencing results. Network studies can illuminate relationships between co-occurring species, and have been used to identify relationships between microbes in contexts such as human microbiomes, agricultural systems, and oceans. This study seeks to build ecological networks of soil fungi for a system with two plant species, Schizachyrium scoparium and Bouteloua curtipendula, sourced from three locations across Texas along a precipitation gradient, for a total of six separate plant – site specific community networks. Soils were sampled in the field and the fungal community was sequenced using ITS2 metabarcoding; OTU data was assigned taxonomy using the UNITE reference database, and ecological guilds were pulled from FunGuild. The OTU table was then subset to include taxa with at least 10 sequencing reads and Spearman correlation was used to identify significant correlations, which were then included in the networks. Network dimensions were calculated and visualized in R and Cytoscape.
The preliminary network, which includes all significant correlations in the dataset, includes 3584 fungal taxa and 50844 significant interactions. It exhibits strong modularity (0.76) with most OTUs interconnected in a central community with 2678 nodes and 39162 edges, which comprises 74.7% of taxa and 77% of significant correlations. The taxa that are not connected to the central community are most likely rare taxa, each present in a single sample. Additional analyses will be done with this preliminary network and with subset networks to compare communities between host plant species and growth sites. FunGuild classifications will be added to identify the roles of functional groups in network cohesion. This study seeks to use network building and analysis to make inferences about belowground fungal-fungal interactions, and how those interactions may be altered by factors such as nearby plant identity, geographical site, and precipitation.
