Improvement in power curves gained by prioritising samples based on genetic risk scores with different predictive powers.

<p>The color of the line represents the proportion of total variance capture by the risk score, with the black line representing a random (i.e. non-prioritised) selection of samples. a) A case-control scenario for a disease with 1% prevalence. The total cohort size for prioritisation is 10,000 cases and an equal number of controls, and we measured power to detect a risk allele with an odds ratio of 2 and a frequency of 1% at genome-wide significance. b) A quantitative trait scenario. The total cohort size is 100,000, and we measured power to detect an allele with 1% frequency that increased a normally-distributed quantitative trait by 0.2 standard deviations. Power is calculated using the trend test power equations given by Freidlin <i>et al </i><a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0076328#pone.0076328-Freidlin1" target="_blank">[36]</a>.