iscc.data.DNABatchHyperParams dataclass

DNABatchHyperParams(breadth: str = 'wgs', depth_model: str = 'dm', mu_depth: float = 30.0, kappa: float = 2000.0, nb_dispersion: float = 0.1, gc_curve_sigma: float = 0.2, mappability_sigma: float = 0.1, capture_sigma: float = 0.0, error_rate: float = 0.001, depth_batch_sigma: float = 0.05, ado_rate: float = 0.0, beta_binom_conc: float = 30.0, doublet_rate: float = 0.0, ffpe_ct_rate: float = 0.0)

The DNA assay's technical parameters — a defaults container.

Holds the protocol-typical magnitudes of the DNA coverage model: the depth emission (capture breadth, mean coverage, and the DM/NB amplification regime), the coverage-bias terms (GC curve, mappability, capture efficiency, depth shift), and the allele layer (sequencing error, single-cell ADO, beta-binomial overdispersion, doublets, FFPE C>T). The scDNA / bulkDNA assays build one and let you override any field, and estimate_dna fits these same parameters from real data.

Attributes:

Name	Type	Description
breadth	str	Capture breadth {wgs, wes, panel} (sets the locus set + depth regime).
depth_model	str	Depth emission {"dm" (default), "nb"}.
mu_depth	float	Mean per-locus coverage (the depth regime; breadth sets it).
kappa	float	DM concentration = amplification regime (large = bulk, small = single-cell).
nb_dispersion	float	NB per-bin overdispersion phi (only for depth_model="nb").
gc_curve_sigma	float	Per-batch GC->coverage curve strength (the GC bias a panel-of-normals fits).
mappability_sigma	float	Unused placeholder kept for symmetry with the genome mappability field.
capture_sigma	float	Per-target (WES) / per-amplicon (panel) capture-efficiency LogNormal sd.
error_rate	float	Per-base sequencing error (false alt/ref on the allele layer).
depth_batch_sigma	float	Per-batch depth-shift LogNormal sd (batches differ in depth).
ado_rate	float	Single-cell allelic dropout Bernoulli prob (one allele lost at a locus).
beta_binom_conc	float	Single-cell allele-fraction overdispersion (Beta-Binomial concentration).
doublet_rate	float	Single-cell doublet fraction.
ffpe_ct_rate	float	Optional FFPE C>T deamination extra-error at C-sites.