Format

Send to

Choose Destination
Ann Hum Genet. 2012 Nov;76(6):484-99. doi: 10.1111/j.1469-1809.2012.00725.x. Epub 2012 Aug 10.

Analysis of secondary phenotype involving the interactive effect of the secondary phenotype and genetic variants on the primary disease.

Author information

1
Department of Biostatistics, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA.

Abstract

A genome-wide association (GWA) study is usually designed as a case-control study, where the presence and absence of the primary disease define the cases and controls, respectively. Using the existing data from GWA studies, investigators are also trying to identify the association between genetic variants and secondary phenotypes, which are defined as traits associated with the primary disease. However, recent studies have shown that bias arises in the estimation of marker-secondary phenotype association using originally collected data. We recently proposed a bias correction approach to accurately estimate the odds ratio (OR) for marker-secondary phenotype association. In this communication, we further investigated whether our bias correction approach is robust for a scenario involving the interactive effect of the secondary phenotype and genetic variants on the primary disease. We found that in such a scenario, our bias correction approach also provides an accurate estimation of OR for marker-secondary phenotype association. We investigated accuracy of our approach using simulation studies and showed that the approach better controlled for type I errors than the existing approaches. We also applied our bias correction approach to the real data analysis of association between an N-acetyltransferase gene, NAT2, and smoking on the basis of colorectal adenoma data.

PMID:
22881407
PMCID:
PMC3472120
DOI:
10.1111/j.1469-1809.2012.00725.x
[Indexed for MEDLINE]
Free PMC Article

Supplemental Content

Full text links

Icon for Wiley Icon for PubMed Central
Loading ...
Support Center