epicarousel.core.Carousel

class epicarousel.core.Carousel(data_name, data_dir, if_bi=1, if_mc_bi=1, threshold=0.0, filter_rate=0.01, chunk_size=10000, carousel_resolution=10, base='/home/metacell/data/metacell/carousel/output', step=4, threads=8, mc_mode='average', index='cell_type', neighbors_method='umap', n_components=50, svd_solver='arpack', shuffle=0, random_state=1)

__init__(data_name, data_dir, if_bi=1, if_mc_bi=1, threshold=0.0, filter_rate=0.01, chunk_size=10000, carousel_resolution=10, base='/home/metacell/data/metacell/carousel/output', step=4, threads=8, mc_mode='average', index='cell_type', neighbors_method='umap', n_components=50, svd_solver='arpack', shuffle=0, random_state=1)

Args:

data_name: Name of data.

data_dir: Path of an h5ad file where the scCAS data count matrix is stored in the compressed sparse row format. AnnData object of shape n_obs × n_vars. Rows correspond to cells and columns to peaks/regions.

if_bi: Whether to binarize the scCAS data count matrix.

if_mc_bi: Whether to binarize the metacell-by-region matrix.

threshold: Threshold for binarizing metacell-by-region matrix.

filter_rate: Proportion for feature selection.

chunk_size: Number of cells in each chunk.

carousel_resolution: Ratio of the number of cells to that of metacells.

base: Export path for EpiCarousel.

step: Length of Walktrap community detection.

threads: Number of parallel processes.

mc_mode: Mode of calculating metacell-by-region matrix.

index: (Optional) Ground truth cell type label of single cells for downstream analysis and evaluation.

Methods

`__init__`(data_name, data_dir[, if_bi, ...])	Args:
`data_split`()	Partition data sequentially into chunks.
`delete_dirs`()	Remove intermediate files.
`identify_metacells`()	Identify metacells.
`make_dirs`()	# Create output directories.
`merge_metacells`()	Aggregate metacells from each chunk.
`metacell_data_clustering`()	Cluster metacells.
`metacell_preprocess`([neighbors_method])	Preprocess metacells.
`result_comparison`()	Evaluation using four clustering strategies.
`shuffle_data`()	Shuffle the initial data.