A unified web cloud computing platform MiMedSurv for microbiome causal mediation analysis with survival responses

Area of Science:

• Microbiome research
• Computational biology
• Statistical genetics

Background:

• Mediation analysis is crucial for understanding microbiome's causal role in linking exposures to diseases.
• Survival analysis is vital in clinical research for tracking disease progression and health outcomes over time.

Purpose of the Study:

• Introduce MiMedSurv, a web platform for microbiome mediation analysis with survival responses.
• Extend existing microbiome analysis tools to incorporate time-to-event data.
• Provide a user-friendly environment for comprehensive microbiome mediation analysis.

Main Methods:

• MiMedSurv performs exploratory non-mediational survival analysis to compare outcomes between exposure groups.
• It identifies mediating roles of the microbiome at both ecosystem (diversity indices) and individual taxa levels.
• The platform supports analysis across various taxonomic hierarchies (phyla to species).

Main Results:

• Demonstrated the utility of MiMedSurv by analyzing the gut microbiome's mediating role between antibiotic treatment and type 1 diabetes.
• The platform successfully identified specific microbial ecosystem and taxa mediating the observed relationship.
• MiMedSurv provides a comprehensive approach to dissecting complex host-microbiome-disease interactions.

Conclusions:

• MiMedSurv offers a powerful and accessible tool for microbiome mediation analysis incorporating survival data.
• The platform facilitates a deeper understanding of how microbial communities influence health trajectories.
• MiMedSurv is freely available, promoting broader application in human microbiome research.