Wyślij emailem: Power estimation and sample size determination for replication studies of genome-wide association studies