Merge pull request #61 from schillerlab/feature/generate_extended_mar…

…ker_table

Add generate_extended_marker_table

sc_toolbox/api/calc/__init__.py

@@ -403,6 +403,52 @@ def correlate_means_to_gene(means: pd.DataFrame, corr_gene: str = "EOMES"):
     return cors
 
 
+def generate_extended_marker_table(
+    adata: AnnData,
+    qval_thresh: float = 0.05,
+    cell_type_label: str = "cell_type",
+    gene_ranks_key: str = "rank_genes_groups",
+):
+    """
+    Generates an extended marker table with cell types and percentages of expressed cell types per cluster.
+    Run scanpy.tl.rank_genes_groups before using this function.
+
+    Args:
+        adata: AnnData object containing ranked genes
+        qval_thresh: Threshold to filter the log fold change for
+        cell_type_label: Label containing all cell types
+        gene_ranks_key: Key for the ranked gene groups (generated by sc.tl.rank_genes_groups)
+
+    Returns:
+        A Pandas DataFrame
+    """
+    result = adata.uns[gene_ranks_key]
+    all_markers = []
+
+    for cluster in result["names"].dtype.names:
+        current = pd.DataFrame(
+            {
+                "gene": result["names"][cluster],
+                "score": result["scores"][cluster],
+                "log_FC": result["logfoldchanges"][cluster],
+                "pval": result["pvals"][cluster],
+                "pval_adj": result["pvals_adj"][cluster],
+                "cell_type": cluster,
+            }
+        )
+
+        # Add percentage expressed per cell type
+        adata.obs["group"] = ["within" if ct == cluster else "outside" for ct in adata.obs.loc[:, cell_type_label]]
+        current = add_percentages(adata, table=current, group_by="group", gene_label="gene", ids=["within", "outside"])
+
+        all_markers.append(current)
+
+    all_markers_df = pd.concat(all_markers)
+    all_markers_df = all_markers[all_markers_df.pval_adj < qval_thresh].copy()
+
+    return all_markers_df
+
+
 def automated_marker_annotation(
     adata: AnnData,
     organism: str,