fit_spide

fit_spide(
    adata,
    condition=None,
    engine='slopes',
    index=None,
    niche=None,
    sigma=None,
    cell_type='cell_type',
    sample_id='sample_id',
    section=None,
    covariates=(),
    depth='loglib',
    offset=None,
    strata=None,
    assay='counts',
    name='Niche',
    min_cells=10,
    min_patients=6,
    min_detect=0.1,
    genes=None,
    tile=None,
    min_tiles=5,
    n_jobs=1,
    verbose=True,
)

Fit spiDE’s per-patient niche models (R fitSpiDE).

Fits, for every index cell type, one negative binomial GLM per gene of its expression on the local density of every other (niche) cell type, with an intercept per patient so that each niche effect is a within-patient slope. The slopes are then tested across patients, pooled or between conditions.

Two engines estimate the between-patient error the tests need. "slopes" (default) fits the shared model once, without the condition, then gives every patient a one-step NB estimate of its own niche slopes with a within-patient spatial sandwich variance; its fit never sees the condition. "sandwich" fits the condition-specific model directly and tests its coefficients with a patient-clustered CR2 sandwich and Bell-McCaffrey degrees of freedom; it needs the condition at fit time.

Parameters

Name Type Description Default
adata AnnData Cells x genes with raw integer counts, obsm["spatial"] and the niche matrices of build_niches. required
condition str The obs column holding a two-level, patient-level condition. The slopes engine only records it; the sandwich engine fits the condition-specific model only when it is given. None
engine ('slopes', 'sandwich') The between-patient error engine. "slopes"
index str or sequence of str Index cell types (default: every type with at least min_cells cells in at least min_patients patients). None
niche str or sequence of str Niche columns to test (default all); the others stay in the model as adjustments. An index type is never tested against its own niche (see merge_niches). None
sigma float The niche bandwidth (needed only if adata carries several). None
cell_type str The obs column of the cell type. "cell_type"
sample_id str The obs column of the patient (the replicated unit). "sample_id"
section str The obs column of the tissue section (default sample_id), used to lay the spatial tiles of the variance estimate. None
covariates str or sequence of str Cell-level obs columns to adjust for (centred; patient-level ones are refused, the patient intercepts absorb them). ()
depth ('loglib', 'nonlinear', 'spatial_spline', 'none') How sequencing depth enters: the centred log library size as a covariate; a 3-df natural spline of it; a library-size spline within each patient that may vary over the section; or not at all (e.g. when offset carries depth). "loglib"
offset str The name of a layer holding a log-scale offset per cell and gene, added to every linear predictor with its coefficient fixed at 1. None
strata str A patient-level obs column (e.g. slide or batch) whose niche slopes are adjusted for in the condition test. Used by the sandwich engine’s condition model here and by the slopes engine at test time; the pooled model never uses it. None
assay str The layer of raw integer counts (X for "counts" when no such layer exists). "counts"
name str The niche obsm prefix. "Niche"
min_cells int Minimum cells of an index type a patient must contribute. 10
min_patients int Minimum patients a type needs. 6
min_detect float Minimum fraction of an index type’s cells in which a gene is detected. 0.1
genes sequence of str Genes to test (default every gene). Library size is always computed from every gene. None
tile float Side of the square tiles of the spatial variance estimate (default 3 x sigma). None
min_tiles int Minimum tiles a patient needs for its own spatial variance. 5
n_jobs int Genes are fitted in blocks over this many workers; the result does not depend on it beyond floating-point summation order. 1
verbose bool Report progress on stderr. True

Returns

Name Type Description
SpiDEFit The fitted models, one entry per index cell type.

See Also

build_niches : Build the niche matrices this reads. test_spide : Test a fit across patients. spide : Build, fit and test in one call.

Examples

from spidepy import build_niches, fit_spide
from spidepy.datasets import toy_spide
adata = build_niches(toy_spide(), sigma=30)
fit = fit_spide(adata, condition="condition", verbose=False)
fit.engine, fit.sigma
('slopes', 30.0)
list(fit.index)
['A', 'B', 'C']
fit.index["A"]["niches"]
['B', 'C']