Adding Unequal Variance to Signature Analysis T-Test (#186)

* make independent t-test have equal variance!

* add tissue balance and rerun separate notebook

* update top features in notebook 2

* scientific data updated the URL!

* add the actual data in case it updates again

* make sure equal variance is false

* remove hard coded best features in visualization

also calculate t-test in figures after filtering tissues

* update full separation plot figure

* rerun full signature pipeline

* make points in figure bigger b/c using test set

* rerun pipeline

* closes #187

* modify figure for rerunning analysis

* rerun pipeline

10.gene-expression-signatures/0.download-validation-data.ipynb

@@ -58,7 +58,7 @@
      "data": {
       "text/plain": [
        "('download/2019-01-22-CellLineSTAR-fpkm-2pass_matrix.txt',\n",
-       " <http.client.HTTPMessage at 0x7fe8cae44320>)"
+       " <http.client.HTTPMessage at 0x7f2d72ae0080>)"
       ]
      },
      "execution_count": 4,
@@ -100,7 +100,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "url = \"https://www.nature.com/articles/sdata201733/tables/3\"\n",
+    "url = \"https://www.nature.com/articles/sdata201733/tables/4\"\n",
     "name = \"nbl_cellline_phenotype.txt\"\n",
     "path = os.path.join(\"download\", name)"
    ]
@@ -223,12 +223,16 @@
     }
    ],
    "source": [
-    "html = requests.get(url).content\n",
+    "if not os.path.isfile(path):\n",
+    "    html = requests.get(url).content\n",
     "\n",
-    "pheno_df = pd.read_html(html)[0]\n",
-    "pheno_df['Cell Line'] = pheno_df['Cell Line'].str.replace(\"-\", \"\")\n",
+    "    pheno_df = pd.read_html(html)[0]\n",
+    "    pheno_df['Cell Line'] = pheno_df['Cell Line'].str.replace(\"-\", \"\")\n",
     "\n",
-    "pheno_df.to_csv(path, sep='\\t', index=False)\n",
+    "    pheno_df.to_csv(path, sep='\\t', index=False)\n",
+    "\n",
+    "else:\n",
+    "    pheno_df = pd.read_csv(path, sep=\"\\t\")\n",
     "\n",
     "pheno_df.head()"
    ]