Network-based multiple sclerosis pathway analysis with GWAS data from 15,000 cases and 30,000 controls

Abstract

Multiple sclerosis (MS) is an inflammatory CNS disease with a substantial genetic component, originally mapped to only the human leukocyte antigen (HLA) region. In the last 5 years, a total of seven genome-wide association studies and one meta-analysis successfully identified 57 non-HLA susceptibility loci. Here, we merged nominal statistical evidence of association and physical evidence of interaction to conduct a protein-interaction-network-based pathway analysis (PINBPA) on two large genetic MS studies comprising a total of 15,317 cases and 29,529 controls. The distribution of nominally significant loci at the gene level matched the patterns of extended linkage disequilibrium in regions of interest. We found that products of genome-wide significantly associated genes are more likely to interact physically and belong to the same or related pathways. We next searched for subnetworks (modules) of genes (and their encoded proteins) enriched with nominally associated loci within each study and identified those modules in common between the two studies. We demonstrate that these modules are more likely to contain genes with bona fide susceptibility variants and, in addition, identify several high-confidence candidates (including BCL10, CD48, REL, TRAF3, and TEC). PINBPA is a powerful approach to gaining further insights into the biology of associated genes and to prioritizing candidates for subsequent genetic studies of complex traits.

Figure 1

Double Manhattan Plot A Manhattan plot showing the gene-level p values of both GWASs used in this study. Gene-level p values from the WTCCC2 GWAS are displayed at the top, and those corresponding to the meta2.5 GWAS are at the bottom. Detailed block structure is shown in an enlarged region in chromosome 1. Blocks were defined as groups of contiguous genes with a p value ≤ 0.05 (grayed area). The individual p value of each gene is displayed as a colored circle ranging from green (not significant) to yellow to red (most significant). The two plots are largely specular, denoting overall replication (see main text).

Figure 2

Connectedness of First-Order Interaction Networks The number of connections among significant genes was evaluated in the background of 1,000 random simulations (see main text). (A) The total number of edges was plotted as a function of the number of significant genes for each study. (B) The size of the largest connected component is plotted as a function of the total number of edges. The colored lines represent the 50^th (green), 90^th (yellow), 95^th (orange), and 99^th (red) percentiles obtained through simulations with random gene sets of similar size.

Figure 3

Intersection Network Of the 118 nodes obtained by the intersection of the resulting networks from each independent study, 88 were arranged in 13 subnetworks (ranging in size from 2 to 27) and 30 nodes remained isolated. Each node represents a gene product, and each edge represents an experimental physical interaction reported in at least two independent publications. Thus, an edge is only displayed if the same interactions were identified in both studies. Isolated nodes in this representation might still have had interactions within each of the studies, but they were not preserved in both. White nodes are not significant. A color scale (yellow to red) denotes the significance of each node in the WTCCC2 study. V-shaped nodes have nominally significant p values in both studies. Nodes with a yellow outline denote genes containing bona fide MS susceptibility variants. Each of the six subnetworks with size ≥ 3 is highlighted by a different background color (subnetworks of size = 2 were grouped under the same background).

Figure 4

Proportion of Validated Discoveries with a Network versus a Nonnetwork Approach Of the 118 genes in the intersection network, 88 genes were arranged in 13 subnetworks of sizes 2–27. Of those, 54 genes were nominally significant in both studies. Fifty-five percent of these genes either were bona fide MS-associated genes (24%) or fell into bona fide MS blocks (31%). Of the 30 singletons from the 118-gene intersection network, 26 had significant p values in both studies. Forty-six percent of these either were bona fide MS-associated genes (11%) or fell into bona fide MS blocks (35%). From the 154 genes with significant p values but not found in networks, only 25% either were bona fide MS-associated genes (8%) or fell into bona fide MS blocks (17%).