Transcriptome data are insufficient to control false discoveries in regulatory network inference

. Author manuscript; available in PMC: 2024 Dec 13.

Published in final edited form as: Cell Syst. 2024 Aug 21;15(8):709–724.e13. doi: 10.1016/j.cels.2024.07.006

Method	Assumptions

Model-X knockoffs⁴⁴	$P (X)$ is known
Gaussian mirror⁵⁹	$P (Y \ \| X)$ is Gaussian
GeneNet⁴³	$P (Y, X)$ is Gaussian
BINCO²⁴	Selection frequencies are U-shaped and the contribution from null hypotheses is decreasing
Permutation⁵³	All features in X are mutually independent