The median survival time of motifs: A novel motif-based indicator reveals the impact of algal blooms on bacterial network interaction stability

Assessing the impact of environmental disturbances, such as harmful algal blooms (HABs), on microbial ecosystem health requires robust ecological indicators. While metrics for global network stability (e.g., natural connectivity) are well-established, indicators capable of quantifying the robustness of micro-scale interaction patterns remain limited. Here, we develop and propose a novel indicator, the Median Survival Time of Motifs (mMST), which integrates network motif theory with survival analysis to quantify the persistence of fundamental interaction subgraphs under perturbation. We applied this indicator, alongside natural connectivity, to assess bacterial networks in particle-attached (AT), free-living (FL), and sediment (SE) habitats of the eutrophic Lake Chaohu during non-bloom and bloom stages. Our results demonstrated that algal blooms significantly reduced both overall network stability and interaction stability in water column communities (AT and FL). Conversely, sediment networks exhibited high robustness across both metrics, likely due to their higher complexity and functional redundancy. Crucially, mMST provided superior mechanistic resolution compared to global metrics, specifically revealing that network destabilization was driven by the disintegration of complex, fully connected subgraphs (e.g., 4-cliques/M6) rather than random edge loss. Consequently, the mMST metric offers a robust, mathematically generalizable tool for diagnosing the micro-structural health of microbial ecosystems under environmental stress. • A novel indicator, Median Survival Time of Motifs (mMST), is developed to quantify the robustness of micro-scale ecological interactions. • Unlike global metrics, mMST offers mechanistic insights by identifying the collapse of specific functional building blocks. • In Lake Chaohu, algal blooms selectively destabilize complex motifs (e.g., M6) in particle-attached bacteria, preceding network failure. • Sediment networks exhibit high mMST values, buffering bacterial interactions against bloom-induced disturbances.