In the simulations, and in our application to the GTEx data, R = 44, so each Uk is a 44 × 44 covariance matrix, and each component of the mixture (eq. 1) is a distribution in 44 dimensions. Visualizing such a distribution is challenging. However, we can get some insight from the first eigenvector of Uk, denoted by vk, which captures the principal direction of the effects. If Uk is dominated by this principal direction, then we can think of effects from that component as being of the form λvk for some scalar λ. For example, if the elements of the vector vk are approximately equal, then component k captures effects that are approximately equal in all conditions; alternatively, if vk has one large element, with other elements close to zero, then component k corresponds to an effect that is strong in only one condition. See Fig. 3 and Supplementary Fig. 2 for illustration.
