By integrating REST-associated silencers and STARR-defined silencers, we annotated nearly 10,000 silencer cCREs, revealing that silencers represent a substantial and previously underappreciated component of regulatory activity. These elements exhibit diverse biochemical, evolutionary, and chromatin properties, indicating that silencers do not form a single homogeneous class. Instead, different experimental and computational strategies capture complementary subsets of repressive regulatory elements, each reflecting distinct mechanisms of gene repression.
Frequently Asked Questions
We compared our silencer cCREs to several previously published human silencer annotations that used distinct experimental strategies. These included:
STARR-seq–based silencers identified by Jayavelu et al.,
FAIRE-based repressive elements from Pang et al.,
Silencers inferred from anti-correlation between chromatin marks and gene expression by Huang and Ovcharenko, and
H3K27me3-rich regions defined by Cai and Fullwood.
Among these, the strongest concordance was with the STARR-seq–based study, reflecting the shared use of functional reporter assays. In contrast, overlap with chromatin-based silencer annotations was more limited, highlighting fundamental differences in how silencers are detected and defined across studies.
Most silencer cCREs identified in this study do not show strong chromatin accessibility in K562 cells, the primary cell line used for functional assays. This contrasts with many earlier approaches that focused on open chromatin regions in the same cell type where silencer activity was measured. Instead, many of our silencer cCREs exhibit accessibility in embryonic tissues or progenitor cell types, suggesting transient accessibility during development. We hypothesize that such transient accessibility enables repressive factors to bind and establish long-term silencing programs that persist after chromatin closure in differentiated cells. As a result, approaches limited to accessible chromatin in mature cell types may systematically miss these elements.
No. Most REST+ and STARR silencer cCREs lack consistent enrichment for classical repressive chromatin marks such as H3K27me3 or bivalent states. This distinguishes them from silencers identified primarily through polycomb-associated chromatin signatures. These differences extend to repeat element overlap and evolutionary conservation, with distinct silencer sets showing enrichment for different repeat classes and conservation profiles. Together, these observations indicate that silencers can operate through multiple epigenetic mechanisms and are not defined by a single chromatin signature.
There is currently no single, universal definition of a silencer. Different studies identify repressive regulatory elements using complementary criteria, including chromatin marks, transcription factor binding, functional reporter assays, or effects on gene expression. Each approach captures elements with different properties and likely different mechanisms of repression. The ENCODE4 Registry provides a unified framework that accommodates this diversity by anchoring silencer discovery within a common cCRE vocabulary. Future work will be needed to further dissect these subclasses, refine definitions, and identify additional silencers that may not yet be captured by existing assays or biosample coverage.