Figure_1.tif (254.36 kB)
Comparison of the statistical power for three categories of prior odds for the null hypothesis when evaluating the noteworthiness of SNPs by BFDP.
figure
posted on 2012-05-25, 00:42 authored by Mattias Johansson, Angus Roberts, Dan Chen, Yaoyong Li, Manon Delahaye-Sourdeix, Niraj Aswani, Mark A. Greenwood, Simone Benhamou, Pagona Lagiou, Ivana Holcátová, Lorenzo Richiardi, Kristina Kjaerheim, Antonio Agudo, Xavier Castellsagué, Tatiana V. Macfarlane, Luigi Barzan, Cristina Canova, Nalin S. Thakker, David I. Conway, Ariana Znaor, Claire M. Healy, Wolfgang Ahrens, David Zaridze, Neonilia Szeszenia-Dabrowska, Jolanta Lissowska, Eleonóra Fabiánová, Ioan Nicolae Mates, Vladimir Bencko, Lenka Foretova, Vladimir Janout, Maria Paula Curado, Sergio Koifman, Ana Menezes, Victor Wünsch-Filho, Jose Eluf-Neto, Paolo Boffetta, Silvia Franceschi, Rolando Herrero, Leticia Fernandez Garrote, Renato Talamini, Stefania Boccia, Pilar Galan, Lars Vatten, Peter Thomson, Diana Zelenika, Mark Lathrop, Graham Byrnes, Hamish Cunningham, Paul Brennan, Jon Wakefield, James D. MckayThese power calculations assume an evaluation of 300,000 SNPs of which 100 are truly associated with the outcome and distributed evenly across three prior categories, respectively. The overall distribution of SNPs across the three prior categories is assumed to be [87.5%; 10%; 2.5%]. Flat PO assumes one single prior category.
