Merge pull request #32 from biosustain/fix_nb

🎨 remove duplication and improve view

Author: enryH

2_data_analysis.ipynb

@@ -552,28 +552,40 @@
    "id": "ae74387d",
    "metadata": {},
    "source": [
-    "# Analytical Plots\n",
-    "- data distribution (e.g. histogram)\n",
+    "# Quality Control Plots (protein level)\n",
+    "- data distribution (histogram and boxplot) to identify drift\n",
     "- coefficient of variation (CV)\n",
     "- number of identified proteins per sample"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "id": "9d7e3e1f",
+   "metadata": {},
+   "source": [
+    "## Data distribution (histogram)"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
    "id": "c2de94d3",
    "metadata": {},
    "outputs": [],
    "source": [
-    "_min_int, _max_int = proteins.min().min(), proteins.max().max()\n",
-    "bins = range(int(_min_int), int(_max_int) + 1, 1)\n",
-    "ax = proteins.T.hist(layout=(2, 4), bins=bins, sharex=True, sharey=True, figsize=(8, 4))\n",
-    "\n",
     "min_int, max_int = int(proteins.min().min()), int(proteins.max().max())\n",
     "bins = range(min_int, max_int+1, 1)\n",
     "ax = proteins.T.hist(layout=(2, 4), bins=bins, sharex=True, sharey=True, figsize=(8, 4))"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "id": "34cba269",
+   "metadata": {},
+   "source": [
+    "## Data distribution (boxplot)"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -596,7 +608,7 @@
    "id": "d39bedbb",
    "metadata": {},
    "source": [
-    "# Coefficient of Variation (CV)\n",
+    "## Coefficient of Variation (CV)\n",
     "- CV = standard deviation / mean\n",
     "  $$ CV = \\frac{\\sigma}{\\mu} $$\n",
     "- per group\n",
@@ -687,7 +699,10 @@
    "source": [
     "# Hierarchical Clustering of normalized data\n",
     "- using completely observed data only\n",
-    "Checkout the [recipe on normalization methods](https://analytics-core.readthedocs.io/latest/api_examples/normalization_analysis.html)."
+    "Checkout the \n",
+    "[recipe on normalization methods](https://analytics-core.readthedocs.io/latest/api_examples/normalization_analysis.html).\n",
+    "\n",
+    "Let's see the effect of normalization on the clustering."
    ]
   },
   {
@@ -1073,7 +1088,7 @@
     "    pval_col=\"padj\",  # toggle if it does not work\n",
     "    correction_alpha=0.2,  # adjust the p-value to see more or less results\n",
     ")\n",
-    "enriched"
+    "enriched.set_index('identifiers')"
    ]
   },
   {
@@ -1153,7 +1168,9 @@
     "    out_dir_subsection / \"1_differently_regulated_as_in_paper.csv\",\n",
     "    index=False,\n",
     ")\n",
-    "view"
+    "view.set_index(\"identifier\")[['pvalue', 'log2FC', 'padj', 'rejected', 'mean(group1)', 'mean(group2)',\n",
+    "       'std(group1)', 'std(group2)',  'test', 'correction', \n",
+    "       'group1', 'group2', 'FC', '-log10 pvalue', 'Method']]"
    ]
   },
   {
@@ -1201,10 +1218,9 @@
     "view = view.set_index(\"identifier\").join(proteins_meta.set_index(\"ProteinName\"))\n",
     "view.to_csv(\n",
     "    out_dir_subsection / \"2_highlighted_proteins_in_figure3.csv\",\n",
-    "    index=False,\n",
+    "    index=True,\n",
     ")\n",
     "sel_cols = [\n",
-    "    \"identifier\",\n",
     "    \"GeneName\",\n",
     "    \"log2FC\",\n",
     "    \"pvalue\",\n",
@@ -1214,7 +1230,7 @@
     "    \"group2\",\n",
     "    \"Method\",\n",
     "]\n",
-    "view.reset_index()[sel_cols].sort_values(\"log2FC\", ascending=False)"
+    "view[sel_cols].sort_values(\"log2FC\", ascending=False)"
    ]
   },
   {

2_data_analysis.py

@@ -315,20 +315,22 @@
 )
 
 # %% [markdown]
-# # Analytical Plots
-# - data distribution (e.g. histogram)
+# # Quality Control Plots (protein level)
+# - data distribution (histogram and boxplot) to identify drift
 # - coefficient of variation (CV)
 # - number of identified proteins per sample
 
-# %%
-_min_int, _max_int = proteins.min().min(), proteins.max().max()
-bins = range(int(_min_int), int(_max_int) + 1, 1)
-ax = proteins.T.hist(layout=(2, 4), bins=bins, sharex=True, sharey=True, figsize=(8, 4))
+# %% [markdown]
+# ## Data distribution (histogram)
 
+# %%
 min_int, max_int = int(proteins.min().min()), int(proteins.max().max())
 bins = range(min_int, max_int+1, 1)
 ax = proteins.T.hist(layout=(2, 4), bins=bins, sharex=True, sharey=True, figsize=(8, 4))
 
+# %% [markdown]
+# ## Data distribution (boxplot)
+
 # %%
 fig, ax = plt.subplots(figsize=(8, 4))
 proteins.T.boxplot(ax=ax)
@@ -340,7 +342,7 @@
 print(f"Saved boxplot to {fname}")
 
 # %% [markdown]
-# # Coefficient of Variation (CV)
+# ## Coefficient of Variation (CV)
 # - CV = standard deviation / mean
 #   $$ CV = \frac{\sigma}{\mu} $$
 # - per group
@@ -396,7 +398,10 @@
 # %% [markdown]
 # # Hierarchical Clustering of normalized data
 # - using completely observed data only
-# Checkout the [recipe on normalization methods](https://analytics-core.readthedocs.io/latest/api_examples/normalization_analysis.html).
+# Checkout the 
+# [recipe on normalization methods](https://analytics-core.readthedocs.io/latest/api_examples/normalization_analysis.html).
+#
+# Let's see the effect of normalization on the clustering.
 
 # %%
 normalization_method = "median"
@@ -617,7 +622,7 @@
     pval_col="padj",  # toggle if it does not work
     correction_alpha=0.2,  # adjust the p-value to see more or less results
 )
-enriched
+enriched.set_index('identifiers')
 
 # %% [markdown]
 # Plot the enrichment scores for the up- and down-regulated proteins separately.
@@ -657,7 +662,9 @@
     out_dir_subsection / "1_differently_regulated_as_in_paper.csv",
     index=False,
 )
-view
+view.set_index("identifier")[['pvalue', 'log2FC', 'padj', 'rejected', 'mean(group1)', 'mean(group2)',
+       'std(group1)', 'std(group2)',  'test', 'correction', 
+       'group1', 'group2', 'FC', '-log10 pvalue', 'Method']]
 
 # %% [markdown]
 # Let's find the proteins highlighted in the volcano plot in Figure 3 in the
@@ -678,10 +685,9 @@
 view = view.set_index("identifier").join(proteins_meta.set_index("ProteinName"))
 view.to_csv(
     out_dir_subsection / "2_highlighted_proteins_in_figure3.csv",
-    index=False,
+    index=True,
 )
 sel_cols = [
-    "identifier",
     "GeneName",
     "log2FC",
     "pvalue",
@@ -691,7 +697,7 @@
     "group2",
     "Method",
 ]
-view.reset_index()[sel_cols].sort_values("log2FC", ascending=False)
+view[sel_cols].sort_values("log2FC", ascending=False)
 
 # %% [markdown]
 # - See normalized data.

material/instructions_statistics_PXD04621.md

@@ -69,7 +69,7 @@ A deployed version of the report can be accessed online at the following link:
 
 ### Download the VueGen HTML report
 
-- [html report](pmultiqc_report/multiqc_report_pmultiqc.html)
+- [html report](./streamlit_report/quarto_report_PXD04621.html)
 
 See instructions on [biosustain/dsp_course_proteomics_intro_report](https://github.com/biosustain/dsp_course_proteomics_intro_report).