scHicCorr 0.0.4

Correlate single-cell Hi-C data.

scHicCorr

Computes per-chromosome correlations between single-cell Hi-C contact matrices stored in .cool files. Supports preprocessing, weighted correlations, and flexible output formats.

Features

• Symmetrizes and normalizes Hi-C contact matrices
• Applies a 2D mean filter to smooth matrices
• Extracts the first K diagonals for analysis
• Computes weighted correlations between multiple datasets
• Outputs compressed CSV (.csv.gz) or Parquet
• Optionally splits results per chromosome
• Parallel processing via multiple CPU cores

Installation

pip install scHicCorr

Usage

schiccorr input1.cool input2.cool --output_prefix results/output --format parquet --split --h 1 --K 5000000 --cores 4

Arguments

Argument	Description	Default
input_files	One or more .cool Hi-C input files
--output_prefix	Prefix for output files (chromosome name appended if --split)
--format	Output format: parquet or csv.gz	parquet
--split	Split output by chromosome into separate files
--h	Mean filter size	1
--K	Number of diagonals to extract	5000000
--cores	Number of CPU cores for parallel processing	1
--log-level	Logging level: DEBUG, INFO, WARNING, ERROR	INFO
--log-file	Write logs to file in addition to stdout

Examples

Single Parquet file:

schiccorr sample1.cool sample2.cool --output_prefix results/hic_corr

Split by chromosome, compressed CSV:

schiccorr sample1.cool sample2.cool --output_prefix results/hic_corr --format csv.gz --split

Log to file:

schiccorr sample1.cool sample2.cool --output_prefix results/hic_corr --log-file run.log

Output

Each row in the output contains reference, comparison, chromosome, and correlation. When --split is used, one file is written per chromosome with the chromosome name appended to the prefix.