A0623
Title: Causal discovery with diverse types of outcomes and unmeasured confounders
Authors: Minjie Wang - Binghamton University (United States) [presenting]
Xiaotong Shen - University of Minnesota (United States)
Wei Pan - University of Minnesota (United States)
Abstract: Causal discovery, the process of identifying causal relationships among variables, is a fundamental problem in statistics. Yet, statistical challenges remain when the data is of mixed data types and affected by unmeasured confounders. These issues are addressed by presenting a novel causal discovery method via instrumental variables with generalized structural equation models suited for analyzing diverse types of outcomes, including discrete, continuous, and mixed data, in the presence of confounders. In particular, two peeling algorithms (bottom-up and top-down) are introduced to ascertain causal relationships and valid instruments. The approach first reconstructs a super-graph to represent ancestral relationships between variables, using a peeling algorithm based on nodewise-constrained GLM regressions that exploit relationships between primary and instrumental variables. Then, it estimates parent-child effects from the ancestral relationships using another peeling algorithm that deconfounds a child's model with information borrowed from its parents' models. A theoretical analysis of the proposed approach is also presented, establishing conditions for model identifiability and providing statistical guarantees for accurately discovering parent-child relationships via the peeling algorithms. Finally, an application to Alzheimer's disease genomics data is demonstrated, highlighting the method's utility in constructing gene-to-gene and gene-to-disease regulatory networks.
