Tested network inference methods.

Methods can be grouped according to the algorithm family or the regularization type. Algorithm families include correlation, partial correlation with inverse covariance, partial covariance with regression, and mutual information. Regularization types employed by methods can be shrinkage, sparsity, or a combination of shrinkage and sparsity as in ELASTICNET. The abbreviations used in this study are given in parentheses in the Method column.