Utilities¶

circe.utils.add_region_infos(adata: AnnData, sep=('_', '_'), inplace=False)¶

Get region informations from the var_names of adata object. e.g. chr1_12345_12346 -> ‘chromosome’ : chr1,

‘start’ : 12345, ‘end’ : 12346

These info will be added to var of anndata object.

adata.var[‘chromosome’] : chromosome adata.var[‘start’] : start position adata.var[‘end’] : end position

Parameters:

  • adata (anndata object) – anndata object with var_names as region names.

  • sep (tuple, optional) – Separator of region names. The default is (‘_’, ‘_’).

Returns:

adata – anndata object with region informations in var.

Return type:

anndata object

circe.utils.cov_to_corr(cov_matrix, tol=1e-20)¶

Optimized version: Convert covariance matrix to correlation matrix, with a tolerance for small diagonal elements.

Parameters:

  • cov_matrix (np.array) – Covariance matrix.

  • tol (float, optional) – Tolerance for diagonal elements. Default is 1e-20.

Returns:

correlation_matrix – Correlation matrix.

Return type:

np.array

Extract upper-triangular links (row < col) from adata.varp[key] into a DataFrame. Works natively with CSR/CSC. If another format is found, it is converted to CSR on the fly (without mutating adata).

Parameters:

  • adata (AnnData) – Object with var_names as variable names.

  • key (str, optional) – Key from adata.varp. If None and only one key exists, that one is used.

  • columns (tuple[str, str, str]) – Output column names (Peak1, Peak2, score).

Returns:

DataFrame with columns given by columns, sorted by descending score.

Return type:

pandas.DataFrame

circe.utils.sort_regions(adata: AnnData)¶

Sort regions by chromosome and start position.

circe.utils.subset_region(adata: AnnData, chromosome, start, end)¶

Subset anndata object on a specific region.

Parameters:

  • adata (anndata object) – anndata object with var_names as region names.

  • chromosome (str) – Chromosome name.

  • start (int) – Start position of the region.

  • end (int) – End position of the region.

Returns:

anndata – anndata object subsetted on the region defined by chr, start and end.

Return type:

anndata object