Eigenvector distributions and optimal shrinkage estimators for large covariance and precision matrices | Synapse