arcfish.tl.TADCaller#

class arcfish.tl.TADCaller(fdr_cutoff: float = 0.1, window: float = 100000.0, tree: bool = True, min_tad_size: float = 0, prominence: float | None = None, distance: int | None = None, method: Literal['pval', 'insulation'] = 'pval')#

TAD calling class.

Parameters:

fdr_cutoff (float, optional) – Boundary peaks with FDR below this cutoff are defined as TAD boundaries, by default 0.1.
window (float, optional) – Domain size in bp to calculate intra- and inter-domain distance, by default 1e5.
tree (bool, optional) – Whether to return hierarchical TADs, by default True.
min_tad_size (float, optional) – The minimum TAD size allowed, by default 0.
prominence (float, optional) – Required only if method is “insulation”, by default None. Least height difference in normalized insulation score in order for the locus to be defined as a peak. Passed to find_peaks().
distance (float, optional) – Required only if method is “insulation”, by default None. Least number of loci between two peaks. Passed to find_peaks().
method (Literal["pval", "insulation"], optional) – TAD calling algorithm used, by default “pval”.

__init__(fdr_cutoff: float = 0.1, window: float = 100000.0, tree: bool = True, min_tad_size: float = 0, prominence: float | None = None, distance: int | None = None, method: Literal['pval', 'insulation'] = 'pval')#

Methods

`__init__`([fdr_cutoff, window, tree, ...])
`by_insulation`(adata)	Call TADs by insulation score.
`by_pval`(adata)	Call TADs by thresholding FDR values.
`call_tads`(adata)	Call TADs from adata.
`to_bedpe`(result)	Convert TAD calling result to a dataframe where each row is a TAD (a pair of boundaries instead of a single boundary).

Attributes

`distance`	Minimum distance between peaks.
`fdr_cutoff`	FDR cut-off for TAD boundaries.
`method`	TAD calling method used.
`min_tad_size`	Minimum TAD size.
`prominence`	Prominence of peaks.
`tree`	Call hierarchical TADs.
`window`	size of the window used to compute inter/intra contacts.

property fdr_cutoff: float#

FDR cut-off for TAD boundaries.

Type:: float

property window: int#

size of the window used to compute inter/intra contacts.

Type:: int

property tree: bool#

Call hierarchical TADs.

Type:: bool

property min_tad_size: int#

Minimum TAD size.

Type:: int

property prominence: float#

Prominence of peaks.

Type:: float

property distance: int#

Minimum distance between peaks.

Type:: int

property method: Literal['pval', 'insulation']#

TAD calling method used.

Type:: str

call_tads(adata: AnnData) → DataFrame#

Call TADs from adata.

Parameters:: adata (AnnData) – adata of a single chromosome, created by arcfish.pp.FOF_CT_Loader.create_adata().
Returns:: A dataframe with length equal to the number of loci. The column peak defines whether the position is a boundary.
Return type:: pd.DataFrame

by_pval(adata: AnnData) → DataFrame#

Call TADs by thresholding FDR values.

Parameters:: adata (AnnData) – adata of a single chromosome, created by arcfish.pp.FOF_CT_Loader.create_adata().
Returns:: A dataframe with length equal to the number of loci. The column peak defines whether the position is a boundary.
Return type:: pd.DataFrame

by_insulation(adata: AnnData) → DataFrame#

Call TADs by insulation score. Method from Su, J.-H., Zheng, P., Kinrot, S. S., Bintu, B. & Zhuang, X. Genome-Scale Imaging of the 3D Organization and Transcriptional Activity of Chromatin. Cell 182, 1641-1659.e26 (2020).

Parameters:: adata (AnnData) – adata of a single chromosome, created by arcfish.pp.FOF_CT_Loader.create_adata().
Returns:: A dataframe with length equal to the number of loci. The column peak defines whether the row is a boundary.
Return type:: pd.DataFrame

to_bedpe(result: DataFrame) → DataFrame | None#

Convert TAD calling result to a dataframe where each row is a TAD (a pair of boundaries instead of a single boundary).

Parameters:: result (pd.DataFrame) – Result returned by by_insulation() or by_pval().
Returns:: Dataframe where each row is a TAD. Columns include c1, s1, e1, c2, s2, e2, {score_col}1, {score_col}2, level, idx1, and idx2, representing the two boundaries of the TAD. If tree is false, then level is 0 for all rows; otherwise, it will be integers increasing from the smaller to larger TADs.
Return type:: pd.DataFrame | None