bbknn.matrix.bbknn
- bbknn.matrix.bbknn(pca, batch_list, neighbors_within_batch=3, n_pcs=50, trim=None, computation='annoy', annoy_n_trees=10, pynndescent_n_neighbors=30, pynndescent_random_state=0, metric='euclidean', set_op_mix_ratio=1, local_connectivity=1, approx=None, use_annoy=None, use_faiss=None, scanpy_logging=False)
Scanpy-independent BBKNN variant that runs on a PCA matrix and list of per-cell batch assignments instead of an AnnData object. Non-data-entry arguments behave the same way as
bbknn.bbknn(). Returns a(distances, connectivities, parameters)tuple, like what would have been stored in the AnnData object. The connectivities are the actual neighbourhood graph.Input
- pca
numpy.array PCA (or other dimensionality reduction) coordinates for each cell, with cells as rows.
- batch_list
numpy.arrayorlist A list of batch assignments for each cell.
- scanpy_logging
bool, optional (default:False) Whether to use scanpy logging to print updates rather than
warnings.warn()
- pca