Accommodating missingness in environmental measurements in gene‐environment interaction analysis

Abstract

For the prognosis of complex diseases, beyond the main effects of genetic (G) and environmental (E) factors, gene-environment (G-E) interactions also play an important role. Many approaches have been developed for detecting important G-E interactions, most of which assume that measurements are complete. In practical data analysis, missingness in E measurements are not uncommon, and failing to properly accommodate such missingness leads to biased estimation and false marker identification. In this study, we conduct G-E interaction analysis with prognosis data under an accelerated failure time (AFT) model. To accommodate missingness in E measurements, we adopt a nonparametric kernel-based data augmentation approach. With a well-designed weighting scheme, a nice “byproduct” is that the proposed approach enjoys a certain robustness property. A penalization approach, which respects the “main effects, interactions” hierarchy, is adopted for selection (of important interactions and main effects) and regularized estimation. The proposed approach has sound interpretations and a solid statistical basis. It outperforms multiple alternatives in simulation. The analysis of TCGA data on lung cancer and melanoma leads to interesting findings and models with superior prediction.

A Selective Review of Multi-Level Omics Data Integration Using Variable Selection

“…It is appealing to conduct integrative G × E analysis to unravel the role that interaction effects play in multi-level omics integration studies and how additional strength that G × E interactions bring to integration. The missingness in environmental and clinical covariates can be efficiently accommodated by the existing approach [127]. Such an endeavor will motivate the development of novel variable selection methods for integrative G × E interactions.…”

Histopathological Imaging–Environment Interactions in Cancer Modeling

“…Studies have shown that tobacco smoking is responsible for 90% of lung cancer [25], and has been identified as a negative prognostic factor for lung adenocarcinoma [26]. In addition, these factors have also been considered in G–E interaction analysis [27].…”

Identification of gene–environment interactions with marginal penalization

“…In particular, the method in Bien et al () is also solved via penalized estimation. With the hierarchy and regularized estimation, the proposed approach has an analysis framework more coherent with that of the joint analysis (Liu et al, ; Wang, Xu, & Ma, ; Wu, Jiang, Ren, Cui, & Ma, , ; Zhao, Zhang, Clark, Maity, & Wu, ). This coherence facilitates methodological developments as well as biological interpretations.…”