Research ArticleECOLOGY

Below-ground plant–fungus network topology is not congruent with above-ground plant–animal network topology

See allHide authors and affiliations

Science Advances  23 Oct 2015:
Vol. 1, no. 9, e1500291
DOI: 10.1126/sciadv.1500291
  • Fig. 1 Plant–fungus network architecture.

    (A) Schematic example of nested and antinested plant–fungus associations. In a nested network, specialists (that is, species with narrow partner ranges) interact with subsets of the partners of generalists (that is, species with broad partner ranges). Networks whose nestedness estimates are higher/lower than that expected by chance are regarded as nested/antinested. Antinestedness can result from compartmentalized and checkerboard network patterns. (B to K) Observed network structure. In each network of (B) cool-temperate (CL) (36 plants and 278 fungi), (C) warm-temperate (WM) (33 plants and 343 fungi), and (D) subtropical (ST) (36 plants and 580 fungi) forests, ectomycorrhizal fungi (red), arbuscular mycorrhizal fungi (blue), and fungi with unknown functions (gray) are linked with their host plants (white). The size of circles represents the relative abundances of fungi or plants in each network. The network-level interaction specialization (E), modularity (F), and nestedness (G) of each network (red bar) are shown with those calculated for randomized networks (blue bar; ±SD). Asterisks represent significant deviations from randomized index values. In addition, checkerboard scores representing how the overlap of host/symbiont plants is avoided in fungal (H) or plant (I) communities are shown. With the use of those network indices, a principal component analysis was also performed (J). The examined ecological networks are plotted on the surface defined by principal components axes (PC1, principal component 1; PC2, principal component 2). For the networks of plants and their root-associated fungi, each index was calculated also for the partial networks representing associations between particular functional or taxonomic groups of fungi and their host plants. The “rank abundance” curve representing the compositional evenness/unevenness of each plant community (K) is also shown.

  • Fig. 2 Partial networks.

    The architecture of partial networks, including each functional or taxonomic group of fungi and their host plants, is shown for the cool-temperate (left), warm-temperate (middle), and subtropical (right) forests. (A to C) Mycorrhizal partial networks including both ectomycorrhizal and arbuscular mycorrhizal fungi. (D to F) Ectomycorrhizal partial networks. (G to I) Arbuscular mycorrhizal partial networks. (J to L) Ascomycete partial networks. (M to O) Basidiomycete partial networks.

  • Fig. 3 Topological properties of partial networks.

    (A to C) The network-level interaction specialization (A), modularity (B), and nestedness (C) of each partial network (red bar) are shown with those calculated for randomized networks (gray bar; ±SD). Asterisks represent significant deviations from randomized index values. MRZ, mycorrhizal partial network; EcM, ectomycorrhizal partial network; AM, arbuscular mycorrhizal partial network; ASC, ascomycete partial network; BSD, basidiomycete partial network.

  • Fig. 4 Network properties explaining variation in (anti)nestedness.

    (A to D) Weighted NODF nestedness (vertical axes) and other network indices (horizontal axes) were calculated for each plant–fungus network or partial network. Positive/negative values indicate that network index estimates of observed plant–fungus association matrices are larger/smaller than those of randomized matrices (that is, relative nestedness, etc.). There was a significant correlation between nestedness and network-level specialization (A) or modularity (B). Further conspicuously, stronger antinestedness was observed for the networks and partial networks with less host-plant overlap in fungal communities (C). The extent of the within-plant-community overlap of fungal symbionts displayed no relationship with network nestedness (D). (E) Multiple regression analysis (df = 10) further indicated that variation in relative nestedness was solely explained by the extent of host range overlap among fungal species.

  • Fig. 5 Contribution of each fungal species.

    (A to C) Fungal species with strong effects on network antinestedness. In each of the cool-temperate (A), warm-temperate (B), and subtropical (C) forests, fungal species that strongly contribute to decreasing network nestedness are indicated in dark blue. White circles represent plant species. (D to F) Relationship between the contribution of each fungal species to network nestedness (vertical axes) and the contribution of each fungal species to reduced host range overlap in a fungal community (horizontal axes). Fungal species that more conspicuously avoided the overlap of host-plant species with others more strongly contributed to the emergence of network antinestedness.

  • Fig. 6 Comparative analysis of factors associated with nestedness.

    Across the metadata of the 59 ecological networks analyzed in Fig. 1J, a multiple regression of nestedness (weighted NODF) was performed with seven network indices (df = 51). Smaller values indicate stronger contribution to antinestedness.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/1/9/e1500291/DC1

    Data File S1. Data sets of the three plant–fungus networks and the list of data sets used in the comparative analysis.

    Fig. S1. Diversity in local networks.

    Fig. S2. Additional analyses of the architectural properties of local plant–fungus networks.

    Fig. S3. Comparison of network architecture between plant–fungus networks and other types of ecological networks.

    Fig. S4. Network modules and nestedness.

    Fig. S5. Relationship between (anti)nestedness and other network architectural properties (r2dtable and vaznull models).

    Fig. S6. Contribution of each fungal species to (anti)nested network architecture.

    Table S1. Randomization tests of network architectural indices with three types of null models.

    Table S2. Randomization tests of checkerboard patterns.

    Table S3. Comparative analysis of factors associated with nestedness (analysis of variance).

  • Supplementary Materials

    This PDF file includes:

    • Fig. S1. Diversity in local networks.
    • Fig. S2. Additional analyses of the architectural properties of local plant–fungus networks.
    • Fig. S3. Comparison of network architecture between plant–fungus networks and other types of ecological networks.
    • Fig. S4. Network modules and nestedness.
    • Fig. S5. Relationship between (anti)nestedness and other network architectural properties (r2dtable and vaznull models).
    • Fig. S6. Contribution of each fungal species to (anti)nested network architecture.
    • Table S1. Randomization tests of network architectural indices with three types of null models.
    • Table S2. Randomization tests of checkerboard patterns.
    • Table S3. Comparative analysis of factors associated with nestedness (analysis of variance).

    Download PDF

    Other Supplementary Material for this manuscript includes the following:

    • Data File S1 (Microsoft Excel format). Data sets of the three plant–fungus networks and the list of data sets used in the comparative analysis.

    Files in this Data Supplement:

Navigate This Article