Computational protocol: Assessing the Power of Exome Chips

Similar protocols

Protocol publication

[…] We provide two types of simulations without adding an effect on the simulated genotypes (0% PAR on all causal variants). First, we aim to characterize the implemented statistical methods with respect to their ability to detect false positive findings. To achieve that, we used the 50 simulated datasets for 60k samples including 100 genes described above and assigned case-control status randomly. For each simulated dataset, we evaluate the percentage of false positives and present the mean percentage across all simulated datasets. In this simulation, we choose a 5% threshold for the p-values of each gene. A genome-wide threshold could have been simulated here as well, but would require a much larger number of null genes and thus dramatically increase the computational burden. Second, we wanted to show the genome-wide performance of the tests with no underlying effect present for the underlying chip structure considered in this paper. Thus, we simulated 10 datasets including all genes (19,975) for two different numbers of individuals (10k, 60k), assigning the case-control status arbitrarily.The simulations and power assessment where done using the computer program R 3.2.1 [], with the additional packages: Matrix [], MultiPhen [] and snpStats [].The simulation program can be received from the authors by request. […]

Pipeline specifications

Software tools MultiPhen, snpStats
Application GWAS