Activity-Informed Network Analysis Reveals Keystone Microbes Shaping Freshwater Ecosystem Function

Freshwater lakes are dynamic ecosystems, with varying oxygen dynamics that influence microbiome structure, composition, and transcriptomic activity. In many freshwater studies, ecological function and abundance metrics are used to discover keystone species; however, it is well established that abundance does not equal activity. Despite the existence of long-term time series spanning multiple years, no previous study has looked at how microbial community and activity (metatranscriptomics) are influenced by shifting oxygen conditions across depths at the microbial network level. In this study, we leverage metagenome-assembled genomes and transcriptomic activity to identify keystone taxa in the ecosystem. Using the SPIEC-EASI and CARlasso methods, we mapped key microbial associations and used permutation-based analyses to assess the robustness of keystone identification. Our results reveal that a taxon's ecological centrality is context-dependent and that many species identified as keystone by abundance alone do not exhibit corresponding transcriptional activity. Notably, members of Bacteroidota and other lineages emerged as keystone taxa only when both abundance and activity were considered. Our study underscores the importance of combining metagenomic and metatranscriptomic approaches for accurate identification of functionally relevant keystone species in freshwater ecosystems, providing a framework for future microbial ecology studies.