scPECA 2.0

PECA2 gene regulatory network construction for single-cell data

scPECA

Introduction

This is a python version of PECA2 gene regulatory network construction software designed for single-cell data. It has a faster running speed and a lower memory footprint.

Installation

pip install scPECA==2.0

Run scPECA

Input

scPECA requires to input the paired (sc)RNA-seq and (sc)ATAC-seq data, and it provides data pre-processing in some formats. (Pre-processing can also be done manually by the user)

Format 1

Paired bulk RNA-seq and ATAC-seq count data

sample_name_RNA.txt

gene1	10
gene2	3

sample_name_ATAC.txt

chr1_10000_10100	1
chr1_20000_20100	0

Format 2

scRNA-seq count data and scATAC-seq count data within the same cluster without meta information

sample_name_scRNA.csv

	barcode1	barcode2
gene1	1	19
gene2	3	6

sample_name_scATAC.csv

	barcode1	barcode2
chr1_10000_10100	1	0
chr1_20000_20100	0	1

Format 3

scRNA-seq count data and scATAC-seq count data with meta information

sample_name_scRNA.csv

	barcode1	barcode2
gene1	1	19
gene2	3	6

sample_name_scRNA_meta.csv

barcode1	celltype1
barcode2	celltype2

sample_name_scATAC.csv

	barcode1	barcode2
chr1_10000_10100	1	0
chr1_20000_20100	0	1

sample_name_scATAC_meta.csv

barcode1	celltype1
barcode2	celltype2

Format 4

h5ad scRNA file with cell label

Prior files

If you are the first time to run scPECA, please download the prior files as follows,

from scPECA.prior_install import figshare_download
import os 
import scPECA
figshare_download(os.path.join(os.path.dirname(scPECA.__file__),'Prior.tar.gz'))

Run example

from scPECA.scPECA_main import scPECAclass
import scPECA
import os

# example 1
pkg_path = os.path.dirname(scPECA.__file__)
# demo data path
data_path = os.path.join(os.path.dirname(scPECA.__file__), 'Cones')
demo = scPECAclass(data_path, 'Cones', 'hg38', pkg_path) # Create a scPECA class
demo.RNA_process(2) # Format 2 RNA data processing
demo.ATAC_process(2) # Format 2 ATAC data processing
demo.network('Cones', data_path) # PECA2 GRN construction

# example 2
data_path = os.path.join(os.path.dirname(scPECA.__file__), '4cellline')
demo = scPECAclass(data_path, '4cellline', 'hg38', pkg_path)
demo.RNA_process(3) 
demo.ATAC_process(3)

# example 3
data_path = os.path.join(os.path.dirname(scPECA.__file__), 'paul15')
demo = scPECAclass(data_path, 'paul15', 'mm10', pkg_path)
demo.RNA_process(4, 'paul15_clusters') 

The details of other optional parameters can be viewed in python. All sample data has been downloaded with the software package.

Output

sample_name_network.txt

TFTG_regulationScore.txt

GRN Analysis Tools (example cell line: K562)

1. TFTG co-module analysis:

demo.co_module(celltype, num_clusters)

Result: TFmodule_result.txt, TGmodule_result.txt, co_module.png

2. TF layering

demo.tf_layer(celltype)
demo.top_tf_layer_plot(celltype)

Result: TF_layer.txt, top_TF_layer_graph.png

System & Software Requirements

System: linux

Linux software: Homer

Python package: pybedtools, ismember, scipy, numpy_groupies,

Considerations

1. RNA data preprocessing currently supports only hg38, hg19, mm10, mm9 genomes.
2. pybedtools may not support the latest version of python and will require the creation of a new environment.