Null-model-based network comparison reveals core associations

Abstract

Microbial network construction and analysis is an important tool in microbial ecology. Such networks are often constructed from statistically inferred associations and may not represent ecological interactions. Hence, microbial association networks are error prone and do not necessarily reflect true community structure. We have developed anuran, a toolbox for investigation of noisy networks with null models. Such models allow researchers to generate data under the null hypothesis that all associations are random, supporting identification of nonrandom patterns in groups of association networks. This toolbox compares multiple networks to identify conserved subsets (core association networks, CANs) and other network properties that are shared across all networks. We apply anuran to a time series of fecal samples from 20 women to demonstrate the existence of CANs in a subset of the sampled individuals. Moreover, we use data from the Global Sponge Project to demonstrate that orders of sponges have a larger CAN than expected at random. In conclusion, this toolbox is a resource for investigators wanting to compare microbial networks across conditions, time series, gradients, or hosts.

Next Generation Microbiome Research: Identification of Keystone Species in the Metabolic Regulation of Host-Gut Microbiota Interplay

“…Finally, recent developments have proposed to apply inferential statistics to network-based models in order to gain confidence in result interpretation. A study by Röttjers et al (2021) recently proposed the use of null models to identify network properties that can be used to compare networks. They showed that among 20 networks built from time-series stool samples collected from 20 women, the new tool called anuran, could identify patterns that were found in 20-25% of the networks but only 3 associations were found in 10 networks or more, suggesting that associations between species or taxa may greatly vary from one individual to another (Röttjers et al, 2021).…”

“…A study by Röttjers et al (2021) recently proposed the use of null models to identify network properties that can be used to compare networks. They showed that among 20 networks built from time-series stool samples collected from 20 women, the new tool called anuran, could identify patterns that were found in 20-25% of the networks but only 3 associations were found in 10 networks or more, suggesting that associations between species or taxa may greatly vary from one individual to another (Röttjers et al, 2021). Although this is a proof-of-concept work applied to a limited population, this is an interesting approach to robustly identify keystone species based on stable interaction networks.…”

Integration of metataxonomic datasets into microbial association networks highlights shared bacterial community dynamics in fermented vegetables

“…By integrating several public datasets together, we were able to characterize two successional shifts that were conserved among different fermentation ecosystems: the first from the initial microbial population of vegetables to Enterobacterales , and the second to an assemblage dominated by Lactobacillales . To test the significance of the core network we obtained, we used an approach based on comparison to a null model, which was similar to that developed by Röttjers et al [22], with a sampling of random graphs similar to Doane et al [32]. Indeed, the identification of core networks is a more challenging task than computation of the global intersection network [21].…”

“…One promising approach for meta-analysis is the construction of microbial association networks, which provide additional and complementary information to classic analyses of alpha- and beta-diversity [18]. Association networks enable the identification of hub species [19, 20], taxa clusters [21], and core networks, the last of which corresponds to the intersection of several microbial association networks and can be used to identify taxa and associations shared by most networks [22]. Association networks were originally designed for macroscopic ecosystems and have only recently been adapted for the investigation of interactions within microbial assemblages [21].…”

Comprehensive non-black box classification of highly correlated ecological time series pairs containing many zeros: the case of gut microbiome of mice

“…To estimate mutual interactions between species, the Lotka–Voltera equation is used as a basic dynamical equation [22,23]. The analysis of linear and nonlinear correlations [24,25] and network analyses [26] is also under intensive investigation. Approaches to estimating causal relationships between species from time series data are also being explored [27,28].…”