remopy

Python implementation of REMO.v1.GRCh38, the R data package from the Stuart Lab.

REMO (Regulatory Element MOdules) provides pre-defined, cell-type annotated regulatory element groupings for single-cell chromatin accessibility analysis.

Installation

# Core data package (just polars)
pip install remopy

# With fragment quantification support
pip install remopy[quantify]

Quick Start

Data Access

import polars as pl
import remopy as remo

# Load module coordinates (1.5M CRE intervals → 340k modules)
modules = remo.modules()
print(modules.head())

# Load module metadata
metadata = remo.metadata()
print(metadata.columns)  # ['REMO', 'CREs', 'Bases', 'Chromosome', 'GC_mean', 'CL']

# Filter modules by cell type: terms() returns cell type name → list of REMO IDs
terms = remo.terms()
t_cell_ids = terms.get('T cell', [])
t_cell_mods = remo.modules().filter(pl.col('REMO').is_in(t_cell_ids))

# Filter by chromosome
chr1_mods = remo.modules().filter(pl.col('chrom') == 'chr1')

# Filter by metadata (e.g. high GC modules)
high_gc_ids = remo.metadata().filter(pl.col('GC_mean') > 0.5)['REMO'].to_list()
high_gc_mods = remo.modules().filter(pl.col('REMO').is_in(high_gc_ids))

# Get cell types present in a tissue
tissues = remo.tissues()
brain_cell_types = tissues.get('Brain', [])

Fragment Quantification (scATAC-seq)

Skip peak calling entirely — quantify fragments into REMO:

import polars as pl
import scanpy as sc
import remopy as remo

# Quantify all modules (requires polars-bio)
adata = remo.quantify('fragments.tsv.gz', min_fragments=1000)

# Or filter modules first — pass a filtered DataFrame to the modules= argument
t_cell_ids = remo.terms().get('T cell', [])
t_cell_mods = remo.modules().filter(pl.col('REMO').is_in(t_cell_ids))
adata = remo.quantify('fragments.tsv.gz', modules=t_cell_mods, min_fragments=1000)

# Standard scanpy workflow
sc.pp.normalize_total(adata, target_sum=1e4)
sc.pp.log1p(adata)
sc.tl.pca(adata)
sc.pp.neighbors(adata)
sc.tl.umap(adata)
sc.tl.leiden(adata)

Data Contents

Data	Description
modules()	1,507,327 CRE intervals grouped into 340,069 modules
metadata()	Module-level stats: CRE count, bases, GC content, cell ontology
terms()	Cell type name → module ID mappings (144 cell types)
ontology()	Cell Ontology ID → module ID mappings
tissues()	Tissue → cell type mappings (25 tissues)

Why REMO?

• No peak calling needed: Use pre-defined, validated features
• Reproducible: Same features across all datasets
• Cell-type annotated: Modules linked to Cell Ontology terms
• Fast: Direct fragment → module quantification

Citation

Please see CITATION.cff for how to cite this package and the REMO paper.

Preprint: Regulatory element modules as universal features for single-cell chromatin analysis