Merge pull request bedapub#273 from bedapub/cell_anno_fix

Small fix in cell annotation workbook as per bedapub#271
idavydov · Feb 7, 2023 · e8237a3 · e8237a3
2 parents 2f1764d + 4d05797
commit e8237a3
Showing 1 changed file with 48 additions and 1 deletion.
diff --git a/workbooks/celltype_annotation_besca.ipynb b/workbooks/celltype_annotation_besca.ipynb
@@ -135,6 +135,15 @@
     "        clusters = \"leiden_citeseq\""
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "adata.uns['log1p'] = {'base' : None} # Fix for bug related to scanpy version scverse/scanpy#2239"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -191,6 +200,24 @@
     ")"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "### Export top marker genes per celltype to .tsv file to facilitate checking the annotation\n",
+    "top5s=list()\n",
+    "for myc in DEgenes.keys():\n",
+    "    top5s=top5s+list(DEgenes[myc].sort_values(\"Log2FC\", ascending=False).iloc[0:5,:]['Name'])\n",
+    "    DEgenes[myc].sort_values(\"Log2FC\", ascending=False).iloc[0:50,:].to_csv(results_folder + \"/figures/TopMarkergenes-\" + clusters + \"_\"+ myc + \"_top50_wilcox.csv\", sep=\"\\t\",index=False)\n",
+    "\n",
+    "### Generate plots of the top marker genes \n",
+    "sc.pl.dotplot(adata, var_names=top5s, groupby=clusters, dot_max=0.8,vmax=3,  dendrogram=True,save=\"TopMarkerplots-\"+clusters+\".svg\")\n",
+    "\n",
+    "sc.pl.matrixplot(adata, var_names=top5s, groupby=clusters, standard_scale='var',  dendrogram=True, save=\"TopMarkerplots-\"+clusters+\".svg\")"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -1332,12 +1359,32 @@
    "metadata": {},
    "outputs": [],
    "source": [
+    "labeloi=\"celltype3\"\n",
+    "\n",
     "### Perform DE cells of each celltype3 vs. all other cells\n",
     "DEgenes = bc.tl.dge.get_de(\n",
-    "    adata, \"celltype3\", demethod=\"wilcoxon\", topnr=5000, logfc=1, padj=0.05\n",
+    "    adata, labeloi, demethod=\"wilcoxon\", topnr=5000, logfc=1, padj=0.05\n",
     ")"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "### Export top marker genes per celltype to .tsv file to facilitate checking the annotation\n",
+    "top5s=list()\n",
+    "for myc in DEgenes.keys():\n",
+    "    top5s=top5s+list(DEgenes[myc].sort_values(\"Log2FC\", ascending=False).iloc[0:5,:]['Name'])\n",
+    "    DEgenes[myc].sort_values(\"Log2FC\", ascending=False).iloc[0:50,:].to_csv(results_folder + \"/figures/TopMarkergenes-\" + labeloi + \"_\"+ myc + \"_top50_wilcox.csv\", sep=\"\\t\",index=False)\n",
+    "\n",
+    "### Generate plots of the top marker genes \n",
+    "sc.pl.dotplot(adata, var_names=top5s, groupby=labeloi, dot_max=0.8,vmax=3, dendrogram=True, save=\"TopMarkerplots-\"+labeloi+\".svg\")\n",
+    "\n",
+    "sc.pl.matrixplot(adata, var_names=top5s, groupby=labeloi, standard_scale='var',  dendrogram=True, save=\"TopMarkerplots-\"+labeloi+\".svg\")"
+   ]
+  },
   {
    "cell_type": "markdown",
    "metadata": {},