 Gaussian conditional independence

For a jointly Gaussian vector, conditional independence is equivalent to vanishing conditional cross-covariance. Equivalently, two coordinates are conditionally independent given all the others exactly when their entry in the precision matrix is zero.