uchrom.strc.loop¶

class uchrom.strc.loop.LoopCallerParams(cut_lo: float = 100000.0, cut_up: float = 1000000.0, inner_cut: float = 25000.0, outer_cut: float = 50000.0, fdr_cutoff: float = 0.1, pval_cutoff: float = 1e-05, gap: float = 50000.0, k_sigma: float = 4.0, frac: float = 0.1, min_cluster_size: int = 1)[source]¶

Bases: object

Runtime parameters for call_loops_axiswise_f().

Defaults mirror ArcFISH’s LoopCaller.

cut_lo: float = 100000.0¶

cut_up: float = 1000000.0¶

fdr_cutoff: float = 0.1¶

frac: float = 0.1¶

gap: float = 50000.0¶

inner_cut: float = 25000.0¶

k_sigma: float = 4.0¶

min_cluster_size: int = 1¶

outer_cut: float = 50000.0¶

pval_cutoff: float = 1e-05¶

uchrom.strc.loop.call_loops_axiswise_f(cd, chrom: str, params: LoopCallerParams | None = None, device: str = 'auto', store: bool = True, result_key: str = 'loops', verbose: bool = False) → DataFrame[source]¶

ArcFISH-style loop caller on a single chromosome.

Parameters:

cd (uchrom.ChromData) – Tracing data. Must contain trace_id in spots and positive coordinates.
chrom (str) – Chromosome name — matched against cd.spots['chrom'].
params (LoopCallerParams, optional) – Override any default.
device (str) – 'auto' | 'cpu' | 'cuda' | 'mps' — passed to GPU-friendly preprocessing.
store (bool) – If True, also write the result DataFrame into cd.results[result_key].
result_key (str) – Key under which to store the result (default 'loops').

Returns:

One row per called loop summit.

Return type:

DataFrame

class uchrom.strc.loop.axiswise_f.LoopCallerParams(cut_lo: float = 100000.0, cut_up: float = 1000000.0, inner_cut: float = 25000.0, outer_cut: float = 50000.0, fdr_cutoff: float = 0.1, pval_cutoff: float = 1e-05, gap: float = 50000.0, k_sigma: float = 4.0, frac: float = 0.1, min_cluster_size: int = 1)[source]¶