Abundant pleiotropy in human complex diseases and traits

Abstract

We present a systematic review of pleiotropy among SNPs and genes reported to show genome-wide association with common complex diseases and traits. We find abundant evidence of pleiotropy; 233 (16.9%) genes and 77 (4.6%) SNPs show pleiotropic effects. SNP pleiotropic status was associated with gene location (p = 0.024; pleiotropic SNPs more often exonic [14.5% versus 4.9% for nonpleiotropic, trait-associated SNPs] and less often intergenic [15.8% versus 23.6%]), "predicted transcript consequence" (p = 0.001; pleiotropic SNPs more often predicted to be structurally deleterious [5% versus 0.4%] but not more often in regulatory sequences), and certain disease classes. We develop a method to calculate the likelihood that pleiotropic links between traits occurred more often than expected and demonstrate that this approach can identify etiological links that are already known (such as between fetal hemoglobin and malaria risk) and those that are not yet established (e.g., between plasma campesterol levels and gallstones risk; and between immunoglobulin A and juvenile idiopathic arthritis). Examples of pleiotropy will accumulate over time, but it is already clear that pleiotropy is a common property of genes and SNPs associated with disease traits, and this will have implications for identification of molecular targets for drug development, future genetic risk-profiling, and classification of diseases.

Figure 1

Interpretation of High Levels of Pleiotropy in Particular Disease Classes If there was a high level of pleiotropy in disease class A, all characteristics of genes associated with disease class A would be associated with pleiotropy (including, but not specifically, the characteristic leading to overrepresentation of disease class A in the pleiotropy category). Therefore, some gene characteristics could be associated with pleiotropy as a result of confounding.

Figure 2

Genes Harboring Variants Associated with Crohn Disease and Other Phenotypes Genes harboring variants associated with other immune-mediated diseases: variants in BACH2, type 1 diabetes and celiac disease; variants in C11orf30 – atopic dermatitis; variants in CCR6 – RA and vitiligo; variants in DENND1B, asthma; variants in ICOSLG, celiac disease; variants in IL10, type 1 diabetes, ulcerative colitis, and Behcet disease; variants in IL12B, psoriasis and psoriatic arthritis; variants in IL18R1, asthma and celiac disease; variants in IL18RAP, celiac disease; variants in IL1RL1, eosinophil count and celiac disease; variants in IL23R, inflammatory bowel disease, psoriasis, ulcerative colitis, ankylosing spondylitis, and Behcet disease; variants in IL27, early-onset inflammatory bowel disease and type 1 diabetes; variants in IL2RA, multiple sclerosis, rheumatoid arthritis, type 1 diabetes, vitiligo, and alopecia areata; variants in MST1, ulcerative colitis and primary sclerosing cholangitis; variants in ORMDL3, asthma, primary biliary cirrhosis, type 1 diabetes, ulcerative colitis, and white blood cell count; variants in PTPN2, type 1 diabetes and celiac disease; variants in PTPN22, rheumatoid arthritis and type 1 diabetes; variants in REL, Hodgkin lymphoma, celiac disease, psoriasis, rheumatoid arthritis, and ulcerative colitis; variants in SMAD3, asthma; variants in STAT3, multiple sclerosis; variants in TAGAP, celiac disease; variants in TYK2, psoriasis and type 1 diabetes; variants in ZMIZ1, breast cancer, celiac disease, and inflammatory bowel disease.