Controlling the genome-wide type I error rate in association mapping experiments
