Fig. 4

Structural differences in inter-kingdom co-occurrence networks exist between infants with typical (inverse) and bacteria or fungi atypical overall alpha diversity trends in the first year of life. Inter-kingdom correlation networks of bacterial and fungal species based on overall alpha diversity trends at A 3 and B 12 months. Infants were classified into overall alpha diversity relationships based on the combination of alpha diversity trends they exhibited for bacteria and fungi. A typical inverse trend was characterized by increasing bacterial and decreasing fungal alpha diversity (n = 50); a bacteria atypical trend was characterized by decreasing bacterial and fungal alpha diversity (n = 21); and a fungi atypical trend was characterized by increasing bacterial and fungal alpha diversity (n = 16). Networks were generated using the fast greedy clustering algorithm with a minimum Pearson correlation coefficient threshold of ± 0.4. Positive correlations are displayed in green and negative correlations in red. Bacterial species are represented by circles and fungal species by triangles. Node size was determined based on degree centrality. Hub taxa are those with the highest betweenness centrality and are labelled with their shape perimeter bolded. Shape colour represents clusters of species more likely to co-occur with one another than with species from outside of these modules. Pair-wise comparisons of network measures were calculated using 5,000 permutations (see Table S4)