Fixing some problems in PCA and VAE sections

dsbook/unsupervised/PCAofCarcinomas.ipynb

@@ -119,7 +119,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "Here we see a non-complet separation of the patients based on their two first eigengenes. This means that the patients gene expression differ and that diference is covered by the first two principal components.\n",
+    "Here we see a non-complete separation of the patients based on their two first sample-specific effects. This means that the patients gene expression differ and that diference is covered by the first two principal components.\n",
     "\n",
     "Lets explore which genes that are most reponsible for that difference. We can do so by investigating their Gene-specific vectors."
    ]
@@ -170,17 +170,10 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "KRT17 seems to be important for both Gene-specific vector 1 and 2. This is possible since [KRT17](https://www.proteinatlas.org/ENSG00000128422-KRT17/pathology) is a known cancer related gene. At this point of the analysis, it is not fully clear what the biological interpretation of PC2 is.  From the eigen gene plot, it is however clear that PC1 seem to capture a large part of the difference beween the LUAD and LUSC samples.\n",
+    "KRT17 seems to be important for both Gene-specific vector 1 and 2. This is possible since [KRT17](https://www.proteinatlas.org/ENSG00000128422-KRT17/pathology) is a known cancer related gene. At this point of the analysis, it is not fully clear what the biological interpretation of PC2 is.  From the sample specific effects plot, it is however clear that PC1 seem to capture a large part of the difference beween the LUAD and LUSC samples.\n",
     "\n",
-    "Note that based on this analysis only can speculate about what makes up this difference. It could be differences in the cancer biology of the tumors, but it could also be batch effects from differences in sample treatments. "
+    "Note that based on this analysis only can speculate about what makes up this difference. The differences could stem from actual differences in the cancer biology of the tumors, but it could equally well be due to technical problems, such as batch effects stemming from differences in sample treatments. "
    ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": []
   }
  ],
  "metadata": {