 |
|
Genetic Association Studies and False Discoveries—Reply
|
|
| Since this article does not have an abstract, we have provided the first 150 words of the full text and any section headings. |
|
|
|
In Reply: The FDR q value is an approach to the problem of multiple comparisons that, although novel in epidemiology, has already become standard for reporting high-throughput genomic studies.1 Dr Horne maintains that the best statistical practice in genetic association studies is correction of P values for multiple comparisons, citing Havill and Dyer.2 However, that article describes 3 approaches to multiple testing, one of which is FDR; none of the 3 is declared "best statistical practice," and the authors state that "which method to use is unclear."
Horne also raises questions about our interpretation of results addressing the association between common variation in 24 clotting-related genes and the incidence of nonfatal venous thrombosis in postmenopausal women. While his concern with replicability of genetic association studies matches our own, Horne appears to misinterpret the statistical analysis presented in our article.
Unlike P values, which express the probability of a false-positive result . . . [Full Text of this Article]
Nicholas L. Smith, PhD, MPH
nlsmith@u.washington.edu
Lucia A. Hindorff, PhD;
Susan R. Heckbert, MD, PhD;
Kenneth Rice, PhD;
Thomas Lumley, PhD
University of Washington
Seattle
Frits R. Rosendaal, MD
Leiden University Medical Center
Leiden, the Netherlands
Bruce M. Psaty, MD, PhD
University of Washington
Seattle
RELATED LETTER
Genetic Association Studies and False Discoveries
Benjamin D. Horne
JAMA. 2007;297(22):2477-2478.
EXTRACT
| FULL TEXT
RELATED ARTICLE
Interpreting Results of Large-Scale Genetic Association Studies: Separating Gold From Fool's Gold
Josée Dupuis and Christopher J. O’Donnell
JAMA. 2007;297(5):529-531.
EXTRACT
| FULL TEXT
|
