Skip to content

DATASET_class.md

Latest commit

 

History

History
75 lines (58 loc) · 3.84 KB

DATASET_class.md

File metadata and controls

75 lines (58 loc) · 3.84 KB
Raw

DATASET_class

DATASET_class is the high-level workflow object. It stores a central parameters dictionary, keeps the active AnnData object on self.adata, and wraps preprocessing, clustering, plotting, and IO functions as chainable methods.

Constructor

sctl.DATASET_class(parameters={}, **kwargs)

The constructor merges default parameters, an optional parameters dictionary, and keyword overrides. It also derives output-directory paths from the active parameters.

Data And Output Methods

  • download_data(**kwargs): download data from download_url.
  • unpack_tar(**kwargs): unpack downloaded tar archives.
  • decompress_downloaded_files(**kwargs): decompress downloaded compressed files.
  • fixtypo_in_downloaded_file_name(**kwargs): rename a downloaded file when needed.
  • make_output_dirs_if_not_exist(): create output directories from parameters.
  • load_data(**kwargs): load data into self.adata.
  • reset_cellxgene_var_names(**kwargs): wrap sctl.pp.reset_cellxgene_var_names.

QC And Filtering Methods

  • basic_filitering(**kwargs): run basic cell and gene count filters.
  • annotate_QC_genes(**kwargs): annotate mitochondrial, ribosomal, hemoglobin, and related QC gene groups.
  • calculate_qc_metrics(**kwargs): compute Scanpy QC metrics.
  • annotate_n_view_adata_raw_counts(**kwargs): annotate QC genes, calculate metrics, and plot raw-count QC summaries.
  • plot_qc_metrics(): plot QC metric summaries.
  • filter_cells_by_anotated_QC_gene(**kwargs): filter cells by annotated QC metrics.
  • remove_genes(**kwargs): remove configured technical gene groups.

Transform Methods

  • norm_log(**kwargs): normalize counts and optionally log transform.
  • HVG_selection_log_norm_seurat(**kwargs): select highly variable genes with the Seurat-style Scanpy workflow.
  • HVG_selection_log_norm_seurat_v3(**kwargs): select highly variable genes with seurat_v3.
  • HVG_removal(**kwargs): subset to highly variable genes.
  • regress_out_anotated_QC_genes(**kwargs): regress out configured QC metrics.
  • scale_func(**kwargs): scale expression values.
  • PCA_func(**kwargs): run PCA.
  • calc_cell_cycle_score(**kwargs): score cell-cycle genes.
  • regress_cell_cycle_score_func(**kwargs): regress out cell-cycle scores.
  • process2scaledPCA(**kwargs): run the combined normalization/HVG/regression/scaling/PCA workflow.

Clustering Methods

  • leiden_clustering(**kwargs): compute neighbors, UMAP, and Leiden clusters.
  • rename_leiden_clusters(**kwargs): map Leiden cluster labels to provided names.
  • leiden_cluster_sil_score(**kwargs): calculate and plot silhouette scores for a Leiden resolution.
  • silhouette_walk_Largest_drop(**kwargs): scan Leiden resolutions and report the largest silhouette-score drop.
  • silhouette_walk_4_Largest_drops(**kwargs): scan Leiden resolutions and report the four largest drops.

Plotting Methods

  • marker_gene_umaps(**kwargs): plot marker-gene UMAPs from configured parameters.
  • marker_gene_umaps_old(...): legacy marker-gene UMAP plotting helper.
  • silhouette_score_n_plot(**kwargs): wrap sctl.pl.silhouette_score_n_plot.
  • silhouette_score_of_obs_key_n_plot(**kwargs): wrap sctl.pl.silhouette_score_of_obs_key_n_plot.
  • plot_batch_obs_key_of_obs_key2(**kwargs): wrap sctl.pl.plot_batch_obs_key_of_obs_key2.
  • plot_percent_obs_key2_per_batch_obs_key(**kwargs): wrap the legacy percent-by-batch plotting helper.
  • plot_adata_row_total_dist(**kwargs): plot row-total distributions.
  • plot_adata_raw_and_X_rowdist(**kwargs): compare raw and active matrix row totals.

Notes

  • Most methods mutate self.adata in place and return self.
  • Parameter names intentionally match the current source, including legacy spellings.
  • Scientific defaults and thresholds are defined in the source class and should be changed only intentionally.