Figure 3
From: Refining multivariate disease phenotypes for high chip heritability
Simulation study: comparison of h2 of commonly used traits with that of the quantitative traits derived by the proposed approach with chosen λ 's. Results from both settings (setting 1 - without covariates, setting 2 - with covariates) are shown. For h2 of individual features in each dataset, we show the highest. Because in our simulation the feature that has the highest h2 estimate is shared across all the datasets with varying number of features, the corresponding h2 curve is a straight line. Since covariates do not affect the h2 estimate when their effect is properly corrected for during the estimation, we do not differentiate the two settings and show one curve for traits derived by equal weighted aggregation of individual features in the data and one for individual feature that has the highest h2 estimate.