Abstract:
|
The human body maintains a close symbiotic relationship with the trillions of microorganisms that live upon and within it. Host gene expression in cooperation with the microbiome has been discovered to play a critical role in disease progression and response. In particular, changes in host gene expression may have a marked impact on the species diversity and abundance. Integration of gene expression and microbiome data can be achieved through mediation modeling. Structural equation modeling has been a popular causal framework, however it maintains strong assumptions about the distribution and association of parameters. We propose a nonparametric approach for selecting significant mediating species for models with high dimensional exposures and mediators. Simulation studies show improved performance compared to traditional mediation methods.
|