cEpiNets Dataset

Genome assembly: GRCh38 / hg38
Coordinate convention: 1-based, closed intervals
Cell type: human microglia and microglia-like cells

Processed Gene–CRE–TF network tables from cEpiNets (context-dependent epigenomic networks), the resource described in:

Fu, T.-T. et al. Context-dependent regulatory networks connect Alzheimer's disease genetics to microglial inflammatory states.

This repository contains the processed Gene–CRE–TF network data associated with cEpiNets. Interactive querying and visualization are available through the cEpiNets browser.

Overview

cEpiNets maps how transcription factors, cis-regulatory elements, and genes connect under different microglial contexts, including AD-related genetic perturbations and inflammatory or environmental conditions.

Each row in the core table represents one condition-specific Gene–CRE–TF association:

Gene — CRE — TF   (within one context)

The cEpiNets atlas described in the manuscript contains 191,292 CREs, 556 non-redundant JASPAR TF motif models, and more than 4 million TF–CRE links across 22 context pairs.

The downloadable table here is the filtered, gene-assigned network used by the public browser and therefore has a different row/CRE count from the full atlas:

gene_cre_tf.tsv.gz	count
rows (Gene–CRE–TF–condition associations)	24,964,295
unique CREs	169,148
unique JASPAR motif models (jaspar_id)	556
unique TF names (tf)	529
unique genes / transcripts	28,837
conditions	22

The 556 jaspar_id values are the paper’s non-redundant JASPAR motif models. Multiple motif models can share the same TF name, so this table has 529 unique tf strings. CRE count is lower than 191,292 because this table retains the top 10% differential CREs after the 100 kb / 3-gene-per-CRE assignment used for the public network.

Data Files

path	role	size
network/gene_cre_tf.tsv.gz	Core Gene–CRE–TF association table, 24,964,295 rows	~121 MB gzipped
metadata/conditions.tsv	The 22 condition values in the network table
metadata/columns.md	Column-level schema and coordinate conventions
LICENSE	CC BY 4.0

Motif-level TF binding sites are not included. They do not contain gene names and are not required to use the network table.

File Schema

See metadata/columns.md. Summary of the core table:

column	example
gene	SORL1
chr	chr20
start	30502691
end	30503583
tf	EHF
jaspar_id	MA0598.4
condition	SORL1KO

The table is self-contained. Neo4j is not required to interpret a row.

Conditions

condition has 22 values, listed in metadata/conditions.tsv. Experimental design for each context is in the manuscript (Table S1).

Genome Assembly

All coordinates are GRCh38 / hg38, 1-based and closed (the start and end of a CRE are included). They are taken from the source chr:start-end interval strings without conversion to BED (0-based, half-open) coordinates.

Usage

import pandas as pd

net = pd.read_csv(
    "network/gene_cre_tf.tsv.gz",
    sep="\t",
    dtype={"start": "int32", "end": "int32"},
)

sorl1 = net[net["gene"] == "SORL1"]
trem2 = net[net["condition"] == "TREM2KO"]

From the Hugging Face Hub:

from huggingface_hub import hf_hub_download
import pandas as pd

path = hf_hub_download(
    repo_id="Thresh514/cEpiNets",
    filename="network/gene_cre_tf.tsv.gz",
    repo_type="dataset",
)
df = pd.read_csv(path, sep="\t")

Citation

Please cite the manuscript (DOI to be added on publication):

Fu, T.-T., Kurkela, M., Tu, J., Zhang, J., Sun, N., Farrer, L., TCW, J., and Hou, L.
Context-dependent regulatory networks connect Alzheimer's disease genetics
to microglial inflammatory states.

https://huggingface.co/datasets/Thresh514/cEpiNets

Correspondence: Lei Hou (leihou@bu.edu) and Julia TCW (juliatcw@bu.edu).

License

This dataset is released under Creative Commons Attribution 4.0 International (CC BY 4.0). See LICENSE.