Large scale meta-analysis characterizes genetic architecture for common psoriasis associated variants

Abstract

Psoriasis is a complex disease of skin with a prevalence of about 2%. We conducted the largest meta-analysis of genome-wide association studies (GWAS) for psoriasis to date, including data from eight different Caucasian cohorts, with a combined effective sample size >39,000 individuals. We identified 16 additional psoriasis susceptibility loci achieving genome-wide significance, increasing the number of identified loci to 63 for European-origin individuals. Functional analysis highlighted the roles of interferon signalling and the NFκB cascade, and we showed that the psoriasis signals are enriched in regulatory elements from different T cells (CD8⁺ T-cells and CD4⁺ T-cells including T_H0, T_H1 and T_H17). The identified loci explain ∼28% of the genetic heritability and generate a discriminatory genetic risk score (AUC=0.76 in our sample) that is significantly correlated with age at onset (p=2 × 10^-89). This study provides a comprehensive layout for the genetic architecture of common variants for psoriasis.

Figure 1. Meta-analysis results.

The ‘Manhattan' plot shows the negative log p values of the meta-analysis. The known loci are coloured in blue; the sixteen novel loci are in red.

Figure 2. Association of psoriasis susceptibility with disease risk.

(a) The effect size (odds ratio, OR) of the risk allele is plotted against the minor allele frequency of the signal among all susceptibility loci. (b) True positive rate versus false positive rate for using genetic risk score to distinguish psoriasis versus control samples. Blue line shows the averaged results among the different cohorts (grey), and the s.e. bars are also shown. (c) The median age-at-onset of psoriasis is plotted against different percentile bins (every 2%) of genetic risk scores for all loci (blue) or all loci without MHC (red).

Figure 3. Biological inference for the psoriasis loci.

(a) Enriched functions (nodes) among the associated loci identified using MEAGA. For illustration purposes, only functions with at least four genes overlapping with other enriched functions are shown (the full list of enriched functions/pathways is shown in Supplementary Table 6). The size of the nodes and the width of the edges correlate with the number of overlapped disease-loci and the number of shared disease-loci, respectively. Nodes with dark blue colour represents higher numbers of overlapped loci while lighter colour represents lower numbers of overlapped loci. The functional annotations for the nodes are presented in Supplementary Fig. 8. (b) The observed-to-expected ratio of the number of regulatory-element overlapped loci versus the enrichment p value. Immune cells are highlighted in blue.