HiCMatrixGenerator


class chr3d.HiCMatrixGenerator(
    chrom_sizes: str,
    assembly: str = 'hg38',
    threads: int = 1,
)

Contact matrix generation using cooler.

Creates .cool and multi-resolution .mcool files from pairs.

Parameters

Parameter	Type	Description
chrom_sizes	`str`	Path to chromosome sizes file
assembly	`str`	Genome assembly name (default: `'hg38'`)
threads	`int`	Number of threads (default: 1)

Methods

create


def create(
    self,
    input_pairs: str,
    output_cool: str,
    resolution: int = 1000,
) -> Dict[str, Any]

Create contact matrix from pairs.

Parameters:

Parameter	Type	Description
input_pairs	`str`	Path to filtered pairs file
output_cool	`str`	Path to output `.cool` file
resolution	`int`	Matrix resolution in bp (default: 1000)

Returns:

Dict[str, Any] with keys:

'output_cool': Path to output .cool file
'cool_size_bytes': Size of cool file
'resolution': Matrix resolution
'num_contacts': Number of contacts

balance


def balance(
    self,
    input_cool: str,
    max_iters: int = 200,
    cis_only: bool = True,
) -> Dict[str, Any]

Balance contact matrix.

Parameters:

Parameter	Type	Description
input_cool	`str`	Path to `.cool` file
max_iters	`int`	Maximum iterations (default: 200)
cis_only	`bool`	Only balance cis contacts (default: `True`)

zoomify


def zoomify(
    self,
    input_cool: str,
    output_mcool: str,
    resolutions: List[int] = None,
) -> Dict[str, Any]

Create multi-resolution .mcool file.

Parameters:

Parameter	Type	Description
input_cool	`str`	Path to `.cool` file
output_mcool	`str`	Path to output `.mcool` file
resolutions	`List[int]`	List of resolutions in bp (default: `[1000, 5000, 10000, 25000, 50000, 100000]`)

Example:


import chr3d as c3d
 
matrix = c3d.HiCMatrixGenerator(
    chrom_sizes="/data/genomes/hg38.chrom.sizes",
    assembly="hg38",
    threads=24
)
 
# Create contact matrix
stats = matrix.create(
    input_pairs="filtered.pairs.gz",
    output_cool="sample.cool"
)
 
# Create multi-resolution matrix
matrix.zoomify(
    input_cool="sample.cool",
    output_mcool="sample.mcool",
    resolutions=[1000, 5000, 10000, 25000, 50000, 100000]
)