POSTm

POSTm is a software tool designed to investigate the relationships between host conditions and microbiome profiles characterized by operational taxonomic units (OTUs). It addresses the challenges of high dimensionality and sparsity in microbiome data by incorporating phylogenetic information in a data-supervised manner. The tool employs a local collapsing test called Phylogeny-Guided Microbiome OTU-Specific Association Test (POST) to make associations between host outcomes and individual OTUs.

Key features of POSTm:
1. Adaptive Information Borrowing: POSTm borrows information from neighboring OTUs in the phylogenetic tree, and the extent of borrowing is guided by phylogenetic distance and the association between the outcome and OTU.
2. Kernel Machine Framework: POSTm operates within the kernel machine framework, allowing it to handle complex OTU effects and extend microbiome tests from the community level to the OTU level.
3. Performance Improvements: In simulation studies, POSTm demonstrates superior performance when the phylogenetic tree provides informative information. When the tree is less informative, it still performs on par with existing methods.
4. Real Data Applications: POSTm has been applied to real-world datasets related to bacterial vaginosis and preterm birth, identifying biologically relevant, outcome-associated OTUs more effectively compared to existing methods.