anndict.adata_dict.build_adata_dict

anndict.adata_dict.build_adata_dict(adata, strata_keys, *, desired_strata=None)

Build a dictionary of AnnData objects split by the obs columns specified in strata_keys.

Parameters:

adata AnnData

The input AnnData.

strata_keys list[str]

List of column names in adata.obs to use for stratification.

desired_strata list | dict | None (default: None)

List of desired strata tuples or a dictionary where keys are the col names supplied in strata_keys, and values are lists of desired strata. If None (Default), all combinations of categories in adata.obs[strata_keys] will be used.

Return type:

AdataDict

Returns:

Flat dictionary of class AdataDict where keys are strata tuples and values are corresponding AnnData subsets.

Raises:

ValueError – If desired_strata is neither a list nor a dictionary of lists.

Examples

Case 1: Take all groups

import pandas as pd
from anndata import AnnData
# Create an example AnnData object
adata = AnnData(obs=pd.DataFrame({
    "Donor": ["Donor1", "Donor1", "Donor2"],
    "Tissue": ["Tissue1", "Tissue2", "Tissue1"]
}))
strata_keys = ["Donor", "Tissue"]
adata_dict = build_adata_dict(adata, strata_keys)
print(adata_dict)
> {
>     ("Donor1", "Tissue1"): adata_d1_t1,
>     ("Donor1", "Tissue2"): adata_d1_t2,
>     ("Donor2", "Tissue1"): adata_d2_t1,
> }

Case 2: Take only some groups

import pandas as pd
from anndata import AnnData
# Create an example AnnData object
adata = AnnData(obs=pd.DataFrame({
    "Donor": ["Donor1", "Donor1", "Donor2"],
    "Tissue": ["Tissue1", "Tissue2", "Tissue1"]
}))
strata_keys = ["Donor", "Tissue"]
desired_strata = {
    "Donor": ["Donor1"], #Take only Donor1
    "Tissue": ["Tissue1", "Tissue2"] #Take Tissue1 and Tissue2
}
adata_dict = build_adata_dict(adata, strata_keys, desired_strata=desired_strata)
print(adata_dict)
> {
>     ("Donor1", "Tissue1"): adata_d1_t1,
>     ("Donor1", "Tissue2"): adata_d1_t2
> }