Abstract/Session Information for Program Number 430T
Session Information
Session Title: Complex Traits: Theory and Methods   Session Type: Poster
Session Location: Exhibit Hall, Level 2, Convention Center   Session Time: Thu 7:00AM-4:30PM
Abstract Information
Program Number: 430T  Presentation Time: Thu, Oct 13, 2011, 3:00PM-4:00PM
Abstract Content
Phenome-wide studies of SNPs from GWAS in a broadly phenotyped population. N. Eriksson, J. Y. Tung, D. A. Hinds, C. B. Do, A. K. Kiefer, B. T. Naughton, J. L. Mountain 23andMe, Inc., Mountain View, CA.

   In many cases, SNPs associated with one trait have been found to be associated with others. This apparent pleiotropy may be due to shared etiology, correlation between the traits, or different causal variants tagged by one SNP. Here we present the results of a systematic search for pleiotropic effects using a broadly phenotyped population of over 65,000 individuals. Participants were drawn from the customer base of 23andMe, Inc. and provided self-reported information on subsets of the approximately 1000 phenotypes, including disease diagnoses, morphological traits, personality tests, and much more. We tested several thousand SNPs (limited to associations found via GWAS as well as rare variants associated with Mendelian disease) against the database of phenotypes. Our approach leads to a number of interesting hypotheses. For example, we find that an HLA SNP originally reported with ulcerative colitis (rs2395185) is associated with type 2 diabetes in our data. SNPs originally associated with tooth development near EDA (rs5936487 and rs4844096) are also associated with other morphological traits, including unibrow, attached earlobes, and lower back hair. The well-known association between FTO and obesity also correlates in our data with choice of sweet versus salty foods. We see evidence for interaction between FTO genotype and food choice with obesity. An association with Crohn's disease near a cluster of interleukin genes (rs2188962) is also linked to severity of mosquito bite reaction in our data. As a final example, we show that a SNP near TERT originally reported with glioma is also associated with skin tags (benign skin growths). Our findings also include many well known facts and novel negative findings. For example, SNPs associated with lung cancer and esophageal cancer influence smoking and drinking frequency in our data. On the other hand, we show that ACTN3 null alleles are not associated with sports choice (or other of our phenotypes) in the general population. Widely studied polymorphisms near DRD2 (rs1800497) and COMT (rs4680) are not significantly associated with any phenotypes. These results are based on a cutoff of 1e-6 for significance and are computed in a group of unrelated individuals of European ancestry controlling for sex, age, and projections onto five principal components. A false-discovery rate analysis yields an estimated false-discovery rate of about 0.025 for these results.

You may contact the first author (during and after the meeting) at

American Society of Human Genetics/ICHG 2011 Meeting Management Office
9650 Rockville Pike, Bethesda, MD
Phone: 301-634-7300, Fax: 301-634-7079
Questions and Comments: ashgmeetings@ashg.org