2017/03/23 by Alexia Jolicoeur‐Martineau, Jolicoeur-Martineau, Alexia, Ashley Wazana +13
Environmental Science · Mathematics · Psychology · #Advanced Causal Inference Techniques #Applications (stat.AP) #Cognitive Abilities and Testing #FOS: Computer and information sciences #Health, Environment, Cognitive Aging
paper · pdf · doi:10.48550/arxiv.1703.08111
openalex publication_date 2017/03/23 · openalex created_date 2022/10/03 · openalex updated_date 2026/07/28
Motivated by the goal of expanding currently existing genotype x environment\ninteraction (GxE) models to simultaneously include multiple genetic variants\nand environmental exposures in a parsimonious way, we developed a novel method\nto estimate the parameters in a GxE model, where G is a weighted sum of genetic\nvariants (genetic score) and E is a weighted sum of environments (environmental\nscore). The approach uses alternating optimization to estimate the parameters\nof the GxE model. This is an iterative process where the genetic score weights,\nthe environmental score weights, and the main model parameters are estimated in\nturn assuming the other parameters to be constant. This technique can be used\nto construct relatively complex interaction models that are constrained to a\nparticular structure, and hence contain fewer parameters.\n We present the model as a two-way interaction longitudinal mixed model, for\nwhich ordinary linear regression is a special case, but it can easily be\nextended to be compatible with k-way interaction models and generalized linear\nmixed models. The model is implemented in R (LEGIT package) and using SAS\nmacros (LEGITSAS). Here we present examples from the Maternal Adversity,\nVulnerability, and Neurodevelopment (MAVAN) study where we improve\nsignificantly upon already existing models using alternating optimization.\nFurthermore, through simulations, we demonstrate the power and validity of this\napproach even with small sample sizes.\n