MiRKAT-MC: A Distance-Based Microbiome Kernel Association Test With Multi-Categorical Outcomes

Area of Science:

• Microbiome research
• Statistical genetics
• Computational biology

Background:

• The microbiome significantly influences human health and disease.
• Clinical studies often involve multi-categorical outcomes (disease subtypes/severity) and clustered data (family/longitudinal studies).
• Existing microbiome association methods struggle with these complex data types.

Purpose of the Study:

• To develop a versatile statistical method for microbiome association analysis with multi-categorical outcomes.
• To accommodate both independent and clustered data structures.
• To incorporate diverse ecological distances for robust association detection.

Main Methods:

• Introduction of the microbiome kernel association test with multi-categorical outcomes (MiRKAT-MC).
• Adaptability to nominal and ordinal outcomes in independent and clustered data.
• Utilization of multiple ecological distances and a pseudo-permutation strategy for significance testing.

Main Results:

• MiRKAT-MC maintains nominal type I error rates across various scenarios.
• The method demonstrates increased statistical power for diverse data types.
• Simulations confirm the robustness and efficiency of MiRKAT-MC.

Conclusions:

• MiRKAT-MC provides a flexible and powerful tool for microbiome association studies with complex outcomes.
• The method offers biological insights when applied to real-world data.
• MiRKAT-MC is accessible via an R package with a GUI.