updated KLIFS pocket notebook for latest kinase schema including KLIF…

…S2UniProt dict
choderalab · Oct 16, 2024 · adc17a1 · adc17a1
1 parent d5b6985
commit adc17a1
Showing 1 changed file with 23 additions and 48 deletions.
diff --git a/notebooks/klifs_pocket.ipynb b/notebooks/klifs_pocket.ipynb
@@ -50,6 +50,20 @@
     "              (j.KLIFS is not None and j.KLIFS.pocket_seq is not None)}"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "cf9e691e-a7aa-4005-aece-ae6183323d4a",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# check that all KLIFS2UniProt alignments working as expected - should be no output\n",
+    "for hgnc, pocket in dict_klifs.items():\n",
+    "    list_temp = [pocket.UniProt.canonical_seq[val-1] == pocket.KLIFS.pocket_seq[idx] \\\n",
+    "                 for idx, (key, val) in enumerate(pocket.KLIFS2UniProt.items()) if val is not None]\n",
+    "    not any(list_temp)"
+   ]
+  },
   {
    "cell_type": "markdown",
    "id": "27513e9e",
@@ -60,7 +74,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 3,
+   "execution_count": 4,
    "id": "3369c424-3727-449a-97c7-1a07114a1541",
    "metadata": {},
    "outputs": [],
@@ -72,7 +86,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 4,
+   "execution_count": 5,
    "id": "2cb0372b-dcbd-4f97-bdff-2045a46e87a8",
    "metadata": {},
    "outputs": [],
@@ -83,7 +97,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 5,
+   "execution_count": 6,
    "id": "88f3481e-6b5a-410f-9061-330820f50f9a",
    "metadata": {},
    "outputs": [],
@@ -96,7 +110,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
+   "execution_count": 7,
    "id": "c61efdee-43bf-47d0-993f-cf5a1bcee962",
    "metadata": {},
    "outputs": [],
@@ -107,7 +121,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 7,
+   "execution_count": 8,
    "id": "7a0a522f-3adb-4d37-9f82-2c5f3b4d8ec0",
    "metadata": {},
    "outputs": [
@@ -236,10 +250,10 @@
     "    dict_annot = colors.DICT_COLORS[str_col][\"DICT_ANNOTATION\"]\n",
     "    if dict_annot is None:\n",
     "        ax.legend(loc=\"upper center\", \n",
-    "              bbox_to_anchor=(0.5, -0.15),\n",
-    "              fancybox=True, \n",
-    "              shadow=True, \n",
-    "              ncol=11)\n",
+    "                  bbox_to_anchor=(0.5, -0.15),\n",
+    "                  fancybox=True, \n",
+    "                  shadow=True, \n",
+    "                  ncol=11)\n",
     "    else:\n",
     "        l1 = ax.legend(loc=\"upper center\", \n",
     "                       bbox_to_anchor=(0.5, -0.15),\n",
@@ -258,45 +272,6 @@
     "    plt.title(f\"{cols} Color Scheme\")\n",
     "    plt.savefig(f\"KLIFS_pocket_{cols}.pdf\", bbox_inches=\"tight\");"
    ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "72a3bec4-20f7-46f2-bcb5-64bded860f72",
-   "metadata": {},
-   "source": [
-    "## Generate KLIFS alignment"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 8,
-   "id": "cd456ee4-5976-412c-bb75-c3bf9f7bc397",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "list_KLIFSPocket = []\n",
-    "\n",
-    "for _, j in dict_klifs.items():\n",
-    "    uniprot_id = j.uniprot_id\n",
-    "    hgnc = j.hgnc_name\n",
-    "    str_uniprot = j.UniProt.canonical_seq\n",
-    "    str_klifs_pocket = j.KLIFS.pocket_seq\n",
-    "    kd_idx = (None, None)        \n",
-    "    # if j.KinCore is not None:\n",
-    "    #     kd_idx = (j.KinCore.start-1, j.KinCore.end-1)\n",
-    "    # else:\n",
-    "    #     kd_idx = (None, None)        \n",
-    "\n",
-    "    dict_temp1 = {\n",
-    "        \"uniprotID\": uniprot_id,\n",
-    "        \"hgncName\": hgnc,\n",
-    "        \"uniprotSeq\": str_uniprot,\n",
-    "        \"klifsSeq\": str_klifs_pocket,\n",
-    "        \"idx_kd\": kd_idx,\n",
-    "        }\n",
-    "\n",
-    "    list_KLIFSPocket.append(klifs.KLIFSPocket(**dict_temp1))"
-   ]
   }
  ],
  "metadata": {