Back

SnakeHichipTF reveals transcription factor logic underlying enhancer-promoter wiring in the human brain

Tan, J.; Wu, Y.; Head, R.; Sun, Y.

2026-04-01 genomics
10.64898/2026.03.30.715320 bioRxiv
Show abstract

Enhancer-promoter interactions are a central feature of gene regulation, yet the regulatory logic that governs their selective formation in complex tissues remains poorly understood. To address this gap, we developed SnakeHichipTF, a reproducible and scalable framework that integrates multi-engine HiChIP analysis with AI-based footprinting to decode the transcription factor (TF) logic underlying enhancer-promoter wiring. Applying SnakeHichipTF to HiChIP datasets from the human Middle Frontal Gyrus (MFG) and Substantia Nigra (SN), we identified distinct region-biased enhancer interactions associated with differential gene expression. MFG-biased interactions were enriched for cognitive and psychiatric associated GWAS traits, whereas SN-biased interactions preferentially intersected lipid and metabolic trait architectures. Integration of TF footprinting revealed that these region-biased interaction networks are governed by distinct TF programs: MFG-biased interactions preferentially recruited TFs linked to neuronal signaling and transcriptional activation, whereas SN-biased interactions were associated with metabolic and stress-responsive regulators. Interestingly, MFG-biased regulatory interactions were significantly enriched for Human Accelerated Regions (HARs), and HAR-associated genes showed elevated expression in humans relative to non-human primates, indicating that cortical enhancer wiring is embedded within evolutionarily modified regulatory elements. Together, by linking 3D chromatin architecture, TF logic, genetic risk, and evolutionary regulatory elements, SnakeHichipTF provides a general framework for dissecting the mechanistic basis of spatial gene regulation.

Matching journals

The top 3 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.