scmags.ScMags

class ScMags(data, labels, gene_ann=None, verbose=True)

A class containing the Count matrix and cluster labels, with methods that can select markers and visualize selected markers

Parameters

• data (Union[ndarray, csr_matrix, csc_matrix]) – Count matrix with rows corresponding to cells and columns to genes in numpy array format.

• labels (ndarray) – One-dimensional numpy array containing cluster labels.

• gene_ann (Optional[ndarray]) – One-dimensional numpy array containing gene annotations.

• verbose (bool) – Verbosity

Returns

Return type

class: ~ScMags.

Examples

>>> import scmags as mg
>>> li = mg.datasets.li()

Methods

__init__(data, labels[, gene_ann, verbose])	Initialize self.
dot_plot([cmap, figsize, log_transform, …])	This function draws the dot plot from the scanpy library In this plot,the colors represent the normalized within-cluster average expression, and the circles represent the within-cluster expression rates.
dp_silh_compute(data, mask, pb_it)	It finds the best combination of given genes by parallel computation based on silhouette index.
filter_genes([in_cls_thres, im_exp, …])	It filters out genes that cannot be cluster-specific markers for each cluster for computational efficiency.
get_filter_genes([ind_return])	Returns the remaining genes after filtering.
get_marker_data([log_norm])	It pulls the selected markers from the count matrix and returns it.
get_markers([ind_return])	Returns the selected markers
knn_classifier([test_ratio, nof_neighbors, …])	This function performs k-NN classification with selected markers.
markers_heatmap([scale, log_norm, cmap, …])	Draws a heatmap with selected markers
markers_tSNE([log_norm, n_iter, perplexity, …])	Draws tSNE plot with selected markers
sel_clust_marker([nof_markers, n_cores, …])	Performs cluster-specific marker selection among filtered genes.
silh_compute(data, mask_labs, i, j)	Calculates silhouette index of any gene

Attributes

get_filt_cluster_thresholds	Returns auto-determined in-cluster expression threshold values
get_filter_gene_scores	Returns the calculated score values of the remaining genes after filtering.