Update interaction plots

BioinfoMachineLearning · Dec 17, 2024 · 840a7fe · 840a7fe
1 parent 8ab777b
commit 840a7fe
Show file tree

Hide file tree

Showing 8 changed files with 216 additions and 48 deletions.
diff --git a/notebooks/astex_method_interaction_analysis_plotting.ipynb b/notebooks/astex_method_interaction_analysis_plotting.ipynb
@@ -510,16 +510,23 @@
     "    with pd.HDFStore(file_path) as store:\n",
     "        for key in store.keys():\n",
     "            for row_index in range(len(store[key])):\n",
+    "                target = store[key].iloc[row_index][\"target\"]\n",
+    "                if not isinstance(target, str):\n",
+    "                    target = target.values[0]\n",
+    "\n",
     "                try:\n",
-    "                    interactions[store[key].iloc[row_index][\"target\"].values[0]].extend(\n",
+    "                    interactions[target].extend(\n",
     "                        [\n",
     "                            f\"{split_string_at_numeric(row[0])[0]}:{split_string_at_numeric(row[1])[0]}:{row[2]}\"\n",
     "                            for row in store[key].iloc[row_index].index.values[:-1]\n",
     "                        ]\n",
     "                    )\n",
     "                except Exception as e:\n",
-    "                    print(f\"Error processing {key} row {row_index} due to: {e}. Skipping...\")\n",
+    "                    print(\n",
+    "                        f\"Error processing {key} row {row_index} for target {target} due to: {e}. Skipping...\"\n",
+    "                    )\n",
     "                    continue\n",
+    "\n",
     "    df_rows = []\n",
     "    for target in interactions:\n",
     "        target_interactions = interactions[target]\n",
@@ -552,7 +559,7 @@
     "\n",
     "assert os.path.exists(\n",
     "    \"astex_diverse_interaction_dataframes.h5\"\n",
-    "), \"No reference Astex Diverse interaction dataframe found.\"\n",
+    "), \"No reference interaction dataframe found.\"\n",
     "reference_df = bin_interactions(\"astex_diverse_interaction_dataframes.h5\", \"Reference\")\n",
     "\n",
     "# combine bins from all method dataframes\n",
@@ -598,16 +605,30 @@
     "            }\n",
     "        )\n",
     "\n",
+    "# plot the EMD and WM values for each method\n",
+    "all_emd_values = [entry[\"EMD\"] for entry in emd_values]\n",
+    "min_emd = np.min(all_emd_values)\n",
+    "max_emd = np.max(all_emd_values)\n",
+    "for entry in emd_values:\n",
+    "    emd = entry[\"EMD\"]\n",
+    "    normalized_score = 1 - (emd - min_emd) / (max_emd - min_emd)\n",
+    "    entry[\"WM\"] = normalized_score\n",
     "\n",
-    "# plot the EMD values for each method\n",
-    "emd_values_df = pd.DataFrame(emd_values, columns=[\"Category\", \"Target\", \"EMD\"])\n",
-    "emd_values_df.to_csv(\"astex_diverse_plif_emd_values.csv\")\n",
+    "emd_values_df = pd.DataFrame(emd_values, columns=[\"Category\", \"Target\", \"EMD\", \"WM\"])\n",
+    "emd_values_df.to_csv(\"astex_diverse_plif_metrics.csv\")\n",
     "\n",
     "plt.figure(figsize=(10, 5))\n",
     "sns.boxplot(data=emd_values_df, x=\"Category\", y=\"EMD\")\n",
     "plt.xlabel(\"\")\n",
     "plt.ylabel(\"PLIF-EMD\")\n",
     "plt.savefig(\"astex_diverse_plif_emd_values.png\")\n",
+    "plt.show()\n",
+    "\n",
+    "plt.figure(figsize=(10, 5))\n",
+    "sns.boxplot(data=emd_values_df, x=\"Category\", y=\"WM\")\n",
+    "plt.xlabel(\"\")\n",
+    "plt.ylabel(\"PLIF-WM\")\n",
+    "plt.savefig(\"astex_diverse_plif_wm_values.png\")\n",
     "plt.show()"
    ]
   }

diff --git a/notebooks/astex_method_interaction_analysis_plotting.py b/notebooks/astex_method_interaction_analysis_plotting.py
@@ -420,16 +420,23 @@ def bin_interactions(file_path, category):
     with pd.HDFStore(file_path) as store:
         for key in store.keys():
             for row_index in range(len(store[key])):
+                target = store[key].iloc[row_index]["target"]
+                if not isinstance(target, str):
+                    target = target.values[0]
+
                 try:
-                    interactions[store[key].iloc[row_index]["target"].values[0]].extend(
+                    interactions[target].extend(
                         [
                             f"{split_string_at_numeric(row[0])[0]}:{split_string_at_numeric(row[1])[0]}:{row[2]}"
                             for row in store[key].iloc[row_index].index.values[:-1]
                         ]
                     )
                 except Exception as e:
-                    print(f"Error processing {key} row {row_index} due to: {e}. Skipping...")
+                    print(
+                        f"Error processing {key} row {row_index} for target {target} due to: {e}. Skipping..."
+                    )
                     continue
+
     df_rows = []
     for target in interactions:
         target_interactions = interactions[target]
@@ -462,7 +469,7 @@ def histogram_to_vector(histogram, bins):
 
 assert os.path.exists(
     "astex_diverse_interaction_dataframes.h5"
-), "No reference Astex Diverse interaction dataframe found."
+), "No reference interaction dataframe found."
 reference_df = bin_interactions("astex_diverse_interaction_dataframes.h5", "Reference")
 
 # combine bins from all method dataframes
@@ -508,14 +515,28 @@ def histogram_to_vector(histogram, bins):
             }
         )
 
+# plot the EMD and WM values for each method
+all_emd_values = [entry["EMD"] for entry in emd_values]
+min_emd = np.min(all_emd_values)
+max_emd = np.max(all_emd_values)
+for entry in emd_values:
+    emd = entry["EMD"]
+    normalized_score = 1 - (emd - min_emd) / (max_emd - min_emd)
+    entry["WM"] = normalized_score
 
-# plot the EMD values for each method
-emd_values_df = pd.DataFrame(emd_values, columns=["Category", "Target", "EMD"])
-emd_values_df.to_csv("astex_diverse_plif_emd_values.csv")
+emd_values_df = pd.DataFrame(emd_values, columns=["Category", "Target", "EMD", "WM"])
+emd_values_df.to_csv("astex_diverse_plif_metrics.csv")
 
 plt.figure(figsize=(10, 5))
 sns.boxplot(data=emd_values_df, x="Category", y="EMD")
 plt.xlabel("")
 plt.ylabel("PLIF-EMD")
 plt.savefig("astex_diverse_plif_emd_values.png")
 plt.show()
+
+plt.figure(figsize=(10, 5))
+sns.boxplot(data=emd_values_df, x="Category", y="WM")
+plt.xlabel("")
+plt.ylabel("PLIF-WM")
+plt.savefig("astex_diverse_plif_wm_values.png")
+plt.show()
diff --git a/notebooks/casp15_method_interaction_analysis_plotting.ipynb b/notebooks/casp15_method_interaction_analysis_plotting.ipynb
@@ -505,16 +505,23 @@
     "    with pd.HDFStore(file_path) as store:\n",
     "        for key in store.keys():\n",
     "            for row_index in range(len(store[key])):\n",
+    "                target = store[key].iloc[row_index][\"target\"]\n",
+    "                if not isinstance(target, str):\n",
+    "                    target = target.values[0]\n",
+    "\n",
     "                try:\n",
-    "                    interactions[store[key].iloc[row_index][\"target\"].values[0]].extend(\n",
+    "                    interactions[target].extend(\n",
     "                        [\n",
     "                            f\"{split_string_at_numeric(row[0])[0]}:{split_string_at_numeric(row[1])[0]}:{row[2]}\"\n",
     "                            for row in store[key].iloc[row_index].index.values[:-1]\n",
     "                        ]\n",
     "                    )\n",
     "                except Exception as e:\n",
-    "                    print(f\"Error processing {key} row {row_index} due to: {e}. Skipping...\")\n",
+    "                    print(\n",
+    "                        f\"Error processing {key} row {row_index} for target {target} due to: {e}. Skipping...\"\n",
+    "                    )\n",
     "                    continue\n",
+    "\n",
     "    df_rows = []\n",
     "    for target in interactions:\n",
     "        target_interactions = interactions[target]\n",
@@ -547,7 +554,7 @@
     "\n",
     "assert os.path.exists(\n",
     "    \"casp15_interaction_dataframes.h5\"\n",
-    "), \"No reference CASP15 interaction dataframe found.\"\n",
+    "), \"No reference interaction dataframe found.\"\n",
     "reference_df = bin_interactions(\"casp15_interaction_dataframes.h5\", \"Reference\")\n",
     "\n",
     "# combine bins from all method dataframes\n",
@@ -593,16 +600,30 @@
     "            }\n",
     "        )\n",
     "\n",
+    "# plot the EMD and WM values for each method\n",
+    "all_emd_values = [entry[\"EMD\"] for entry in emd_values]\n",
+    "min_emd = np.min(all_emd_values)\n",
+    "max_emd = np.max(all_emd_values)\n",
+    "for entry in emd_values:\n",
+    "    emd = entry[\"EMD\"]\n",
+    "    normalized_score = 1 - (emd - min_emd) / (max_emd - min_emd)\n",
+    "    entry[\"WM\"] = normalized_score\n",
     "\n",
-    "# plot the EMD values for each method\n",
-    "emd_values_df = pd.DataFrame(emd_values, columns=[\"Category\", \"Target\", \"EMD\"])\n",
-    "emd_values_df.to_csv(\"casp15_plif_emd_values.csv\")\n",
+    "emd_values_df = pd.DataFrame(emd_values, columns=[\"Category\", \"Target\", \"EMD\", \"WM\"])\n",
+    "emd_values_df.to_csv(\"casp15_plif_metrics.csv\")\n",
     "\n",
     "plt.figure(figsize=(10, 5))\n",
     "sns.boxplot(data=emd_values_df, x=\"Category\", y=\"EMD\")\n",
     "plt.xlabel(\"\")\n",
     "plt.ylabel(\"PLIF-EMD\")\n",
     "plt.savefig(\"casp15_plif_emd_values.png\")\n",
+    "plt.show()\n",
+    "\n",
+    "plt.figure(figsize=(10, 5))\n",
+    "sns.boxplot(data=emd_values_df, x=\"Category\", y=\"WM\")\n",
+    "plt.xlabel(\"\")\n",
+    "plt.ylabel(\"PLIF-WM\")\n",
+    "plt.savefig(\"casp15_plif_wm_values.png\")\n",
     "plt.show()"
    ]
   }

diff --git a/notebooks/casp15_method_interaction_analysis_plotting.py b/notebooks/casp15_method_interaction_analysis_plotting.py
@@ -415,16 +415,23 @@ def bin_interactions(file_path, category):
     with pd.HDFStore(file_path) as store:
         for key in store.keys():
             for row_index in range(len(store[key])):
+                target = store[key].iloc[row_index]["target"]
+                if not isinstance(target, str):
+                    target = target.values[0]
+
                 try:
-                    interactions[store[key].iloc[row_index]["target"].values[0]].extend(
+                    interactions[target].extend(
                         [
                             f"{split_string_at_numeric(row[0])[0]}:{split_string_at_numeric(row[1])[0]}:{row[2]}"
                             for row in store[key].iloc[row_index].index.values[:-1]
                         ]
                     )
                 except Exception as e:
-                    print(f"Error processing {key} row {row_index} due to: {e}. Skipping...")
+                    print(
+                        f"Error processing {key} row {row_index} for target {target} due to: {e}. Skipping..."
+                    )
                     continue
+
     df_rows = []
     for target in interactions:
         target_interactions = interactions[target]
@@ -457,7 +464,7 @@ def histogram_to_vector(histogram, bins):
 
 assert os.path.exists(
     "casp15_interaction_dataframes.h5"
-), "No reference CASP15 interaction dataframe found."
+), "No reference interaction dataframe found."
 reference_df = bin_interactions("casp15_interaction_dataframes.h5", "Reference")
 
 # combine bins from all method dataframes
@@ -503,14 +510,28 @@ def histogram_to_vector(histogram, bins):
             }
         )
 
+# plot the EMD and WM values for each method
+all_emd_values = [entry["EMD"] for entry in emd_values]
+min_emd = np.min(all_emd_values)
+max_emd = np.max(all_emd_values)
+for entry in emd_values:
+    emd = entry["EMD"]
+    normalized_score = 1 - (emd - min_emd) / (max_emd - min_emd)
+    entry["WM"] = normalized_score
 
-# plot the EMD values for each method
-emd_values_df = pd.DataFrame(emd_values, columns=["Category", "Target", "EMD"])
-emd_values_df.to_csv("casp15_plif_emd_values.csv")
+emd_values_df = pd.DataFrame(emd_values, columns=["Category", "Target", "EMD", "WM"])
+emd_values_df.to_csv("casp15_plif_metrics.csv")
 
 plt.figure(figsize=(10, 5))
 sns.boxplot(data=emd_values_df, x="Category", y="EMD")
 plt.xlabel("")
 plt.ylabel("PLIF-EMD")
 plt.savefig("casp15_plif_emd_values.png")
 plt.show()
+
+plt.figure(figsize=(10, 5))
+sns.boxplot(data=emd_values_df, x="Category", y="WM")
+plt.xlabel("")
+plt.ylabel("PLIF-WM")
+plt.savefig("casp15_plif_wm_values.png")
+plt.show()
diff --git a/notebooks/dockgen_method_interaction_analysis_plotting.ipynb b/notebooks/dockgen_method_interaction_analysis_plotting.ipynb
@@ -523,16 +523,23 @@
     "    with pd.HDFStore(file_path) as store:\n",
     "        for key in store.keys():\n",
     "            for row_index in range(len(store[key])):\n",
+    "                target = store[key].iloc[row_index][\"target\"]\n",
+    "                if not isinstance(target, str):\n",
+    "                    target = target.values[0]\n",
+    "\n",
     "                try:\n",
-    "                    interactions[store[key].iloc[row_index][\"target\"].values[0]].extend(\n",
+    "                    interactions[target].extend(\n",
     "                        [\n",
     "                            f\"{split_string_at_numeric(row[0])[0]}:{split_string_at_numeric(row[1])[0]}:{row[2]}\"\n",
     "                            for row in store[key].iloc[row_index].index.values[:-1]\n",
     "                        ]\n",
     "                    )\n",
     "                except Exception as e:\n",
-    "                    print(f\"Error processing {key} row {row_index} due to: {e}. Skipping...\")\n",
+    "                    print(\n",
+    "                        f\"Error processing {key} row {row_index} for target {target} due to: {e}. Skipping...\"\n",
+    "                    )\n",
     "                    continue\n",
+    "\n",
     "    df_rows = []\n",
     "    for target in interactions:\n",
     "        target_interactions = interactions[target]\n",
@@ -565,7 +572,7 @@
     "\n",
     "assert os.path.exists(\n",
     "    \"dockgen_interaction_dataframes.h5\"\n",
-    "), \"No reference DockGen interaction dataframe found.\"\n",
+    "), \"No reference interaction dataframe found.\"\n",
     "reference_df = bin_interactions(\"dockgen_interaction_dataframes.h5\", \"Reference\")\n",
     "\n",
     "# combine bins from all method dataframes\n",
@@ -611,16 +618,30 @@
     "            }\n",
     "        )\n",
     "\n",
+    "# plot the EMD and WM values for each method\n",
+    "all_emd_values = [entry[\"EMD\"] for entry in emd_values]\n",
+    "min_emd = np.min(all_emd_values)\n",
+    "max_emd = np.max(all_emd_values)\n",
+    "for entry in emd_values:\n",
+    "    emd = entry[\"EMD\"]\n",
+    "    normalized_score = 1 - (emd - min_emd) / (max_emd - min_emd)\n",
+    "    entry[\"WM\"] = normalized_score\n",
     "\n",
-    "# plot the EMD values for each method\n",
-    "emd_values_df = pd.DataFrame(emd_values, columns=[\"Category\", \"Target\", \"EMD\"])\n",
-    "emd_values_df.to_csv(\"dockgen_plif_emd_values.csv\")\n",
+    "emd_values_df = pd.DataFrame(emd_values, columns=[\"Category\", \"Target\", \"EMD\", \"WM\"])\n",
+    "emd_values_df.to_csv(\"dockgen_plif_metrics.csv\")\n",
     "\n",
     "plt.figure(figsize=(10, 5))\n",
     "sns.boxplot(data=emd_values_df, x=\"Category\", y=\"EMD\")\n",
     "plt.xlabel(\"\")\n",
     "plt.ylabel(\"PLIF-EMD\")\n",
     "plt.savefig(\"dockgen_plif_emd_values.png\")\n",
+    "plt.show()\n",
+    "\n",
+    "plt.figure(figsize=(10, 5))\n",
+    "sns.boxplot(data=emd_values_df, x=\"Category\", y=\"WM\")\n",
+    "plt.xlabel(\"\")\n",
+    "plt.ylabel(\"PLIF-WM\")\n",
+    "plt.savefig(\"dockgen_plif_wm_values.png\")\n",
     "plt.show()"
    ]
   }