circe-py 0.2.2

Circe: Package for building co-accessibility networks from ATAC-seq data.

---

CIRCE: Cis-regulatory interactions between chromatin regions

Description

This repo contains a python package for inferring co-accessibility networks from single-cell ATAC-seq data, using skggm for the graphical lasso and scanpy for data processing.

It is based on the pipeline and hypotheses presented in the manuscript "Cicero Predicts cis-Regulatory DNA Interactions from Single-Cell Chromatin Accessibility Data" by Pliner et al. (2018). This R package Cicero is available here.

Results may slitghly vary between both packages, notably due to the different implementations of graphical lasso.
Currently, scores are very close when applied to the same metacells, computed from Cicero's methodology. (cf comparison plots below). It should run significantly faster than Cicero (e.g.: running time of 5 sec instead of 17 min for the dataset 2).

If you have any suggestion, don't hesitate ! This package is still a work in progress :)

Installation

The package can be installed using pip:

pip install circe-py

and from github

pip install "git+https://github.com/cantinilab/circe.git"

Warning: If you clone the repo, don't stay in the repo to run your script because python will import the non-compiled cython file (probable error: circe.pyquic does not have a quic function)

Minimal example

import anndata as ad
import circe as ci

atac = ad.read_h5ad('atac_data.h5ad')
atac = ci.add_region_infos(atac)
ci.compute_atac_network(atac)
df_network = ci.extract_atac_links(atac)

Comparison to Cicero R package

On the same metacells obtained from Cicero code.

All tests can be found in the circe benchmark repo

Toy dataset 1 (fake data):

• Pearson correlation coefficient: 0.999126
• Spearman correlation coefficient: 0.99838

Real dataset 2 (subsample of neurips PBMC)

• Pearson correlation coefficient: 0.999958
• Spearman correlation coefficient: 0.999911

Performance on real dataset 2:

• Runtime: ~100x faster
• Memory usage: ~5x less

Coming:

• Calculate metacells !
• Add stats on similarity on large datasets.
• Add stats on runtime, memory usage.
• Implement the multithreading use. Should speed up even more.
• Fix seed for reproducibility.

Usage

It is currently developped to work with AnnData objects. Check Example1.ipynb for a simple usage example.