JAM unites both ideas of mediation and latent clustering using summary statistics from multiple omic studies and develops a causal inferential framework to identify mediating effects of biologically relevant factors on outcomes. Importantly, the approach is robust to many of the underlying assumptions made by current MR approaches, such as adjusting for pleiotropic effects by including and selecting a large number of mediators jointly. The approach is flexible and can be extended o high-demisional data and multiethnic or muli-tissue summary statistics. Moreover, the approach can identify shared biological pathways by estimating latent combinations of the intermediates associated with the outcome. Using only summary statistics, this approach is innovative by going well beyond current methods to characterize pathways and corresponding intermediates and SNPs contributing to those associations.
Selected Related Publications
Sylvia Shen (shenjiay at usc dot edu)