Pattern Discovery in Brain Imaging Genetics via SCCA Modeling with a Generic Non-convex Penalty

Abstract

Brain imaging genetics intends to uncover associations between genetic markers and neuroimaging quantitative traits. Sparse canonical correlation analysis (SCCA) can discover bi-multivariate associations and select relevant features, and is becoming popular in imaging genetic studies. The L1-norm function is not only convex, but also singular at the origin, which is a necessary condition for sparsity. Thus most SCCA methods impose -norm onto the individual feature or the structure level of features to pursuit corresponding sparsity. However, the -norm penalty over-penalizes large coefficients and may incurs estimation bias. A number of non-convex penalties are proposed to reduce the estimation bias in regression tasks. But using them in SCCA remains largely unexplored. In this paper, we design a unified non-convex SCCA model, based on seven non-convex functions, for unbiased estimation and stable feature selection simultaneously. We also propose an efficient optimization algorithm. The proposed method obtains both higher correlation coefficients and better canonical loading patterns. Specifically, these SCCA methods with non-convex penalties discover a strong association between the APOE e4 rs429358 SNP and the hippocampus region of the brain. They both are Alzheimer’s disease related biomarkers, indicating the potential and power of the non-convex methods in brain imaging genetics.

Image Genetic Analysis and Application Research Based on QRFPR and Other Neural Network-Related SNP Loci

“…The L1-norm function is not only convex at the origin but also singular, which is a necessary condition for sparsity. Therefore, most SCCA methods will impose individual features or structural hierarchy of features in pursuit of corresponding sparsity [ 2 ]. The research was not objective, and the environmental factors in which the research occurred were not taken into account.…”

Brain Imaging Genomics: Integrated Analysis and Machine Learning

“…Du et al [171] proposed an SCCA framework using a generic nonconvex penalty (GNC-SCCA) to address the challenge that the l1-norm overpenalizes large coefficients and may introduce estimation bias. They tested seven nonconvex penalties for replacing the l1 term in an l1-based SCCA.…”