Detecting genetic associations with brain imaging phenotypes in Alzheimer's disease via a novel structured SCCA approach

Abstract

Brain imaging genetics becomes an important research topic since it can reveal complex associations between genetic factors and the structures or functions of the human brain. Sparse canonical correlation analysis (SCCA) is a popular bi-multivariate association identification method. To mine the complex genetic basis of brain imaging phenotypes, there arise many SCCA methods with a variety of norms for incorporating different structures of interest. They often use the group lasso penalty, the fused lasso or the graph/network guided fused lasso ones. However, the group lasso methods have limited capability because of the incomplete or unavailable prior knowledge in real applications. The fused lasso and graph/network guided methods are sensitive to the sign of the sample correlation which may be incorrectly estimated. In this paper, we introduce two new penalties to improve the fused lasso and the graph/network guided lasso penalties in structured sparse learning. We impose both penalties to the SCCA model and propose an optimization algorithm to solve it. The proposed SCCA method has a strong upper bound of grouping effects for both positively and negatively highly correlated variables. We show that, on both synthetic and real neuroimaging genetics data, the proposed SCCA method performs better than or equally to the conventional methods using fused lasso or graph/network guided fused lasso. In particular, the proposed method identifies higher canonical correlation coefficients and captures clearer canonical weight patterns, demonstrating its promising capability in revealing biologically meaningful imaging genetic associations.

Keywords: Brain imaging genetics; Fused pairwise group Lasso; Graph guided pairwise group Lasso; Sparse canonical correlation analysis (SCCA).

Figure 1:

Canonical weights estimated on synthetic data. The first row is the ground truth, and each remaining row corresponds to an SCCA method: (1) FL-SCCA, (2) NS-SCCA, (3) AGN-SCCA, and (4) FGL-SCCA. For each method, the estimated weights of u are shown on the left panel, and those of v are shown on the right. In each subfigure, the vertical axis represents the indices of each u (left panel) or v (right panel), and the horizontal axis represents 250 runs of experiments (50 times of 5-fold cross-validation).

Figure 2:

Canonical weights estimated on real imaging genetics data set. Each row corresponds to a method: (1) FL-SCCA, (2) NS-SCCA, (3) AGN-SCCA, and (4) FGL-SCCA. For each method, the estimated weights of u are shown on the left panel, and those of v are shown on the right. In each subfigure, the horizontal axis represents the reference number of each individual SNP (left panel) or imaging ROI (right panel), and the vertical axis represents every run and there are 250 runs in total (50 times of 5-fold cross-validation).

Figure 3:

Mapping averaged canonical weights v of FGL-SCCA onto the brain.

Figure 4:

Heat map of brain ROI-SNP associations of top selected markers.

Figure 5:

Heat map of brain ROI-SNP associations of top selected markers.

Figure 6:

Pairwise comparisons in terms of genotype of rs769450 and rs429358 within ADs, MCIs and HCs respectively. Two-way ANOVA was applied to examine the effects of rs769450 and baseline diagnosis on left middle frontal gyrus (a). Age, gender, education, handedness were included as covariates. The results of rs429358 were also shown for comparison (b).