The sub-module catalogs Conserved Non-coding Elements (CNEs) derived from a whole-genome multiple alignment of 21 mammalian species generated using Cactus. To construct this dataset, highly conserved genomic regions were initially identified from the cross-species alignments using the PHAST package. Subsequently, these regions were rigorously filtered based on the reference genome annotations of the human and the pig to exclusively retain non-coding sequences, which serve as a curated resource for investigating evolutionarily constrained regulatory regions across mammals.