make smiles canonical

binding_affinity.py

@@ -90,7 +90,7 @@ class BindingAffinity(datasets.ArrowBasedBuilder):
         features = datasets.Features(
             {
                 "seq": datasets.Value("string"),
-                "smiles": datasets.Value("string"),
+                "smiles_can": datasets.Value("string"),
                 "neg_log10_affinity_M": datasets.Value("float"),
                 "affinity": datasets.Value("float"),
                 # These are the features of your dataset like images, labels ...
@@ -161,7 +161,7 @@ class BindingAffinity(datasets.ArrowBasedBuilder):
         for k, row in df.iterrows():
             yield k, {
                 "seq": row["seq"],
-                "smiles": row["smiles"],
+                "smiles_can": row["smiles_can"],
                 "neg_log10_affinity_M": row["neg_log10_affinity_M"],
                 "affinity_uM": row["affinity_uM"],
             }

combine_dbs.ipynb

@@ -2,7 +2,7 @@
  "cells": [
   {
    "cell_type": "code",
-   "execution_count": 2,
+   "execution_count": 1,
    "id": "95bd761a-fe51-4a8e-bc70-1365260ba5f8",
    "metadata": {},
    "outputs": [],
@@ -1363,17 +1363,62 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "df.to_parquet('data/all.parquet')"
+    "df.to_parquet('data/all_nocan.parquet')"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 3,
+   "execution_count": 11,
    "id": "4e2d89f7-f6ea-41de-a13b-4a184b4fd580",
    "metadata": {},
    "outputs": [],
    "source": [
-    "df = pd.read_parquet('data/all.parquet')"
+    "df = pd.read_parquet('data/all_nocan.parquet')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "b4b9acd7-7784-492b-9fa3-b7fad9d18a9d",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "INFO: Pandarallel will run on 256 workers.\n",
+      "INFO: Pandarallel will use Memory file system to transfer data between the main process and workers.\n"
+     ]
+    }
+   ],
+   "source": [
+    "from pandarallel import pandarallel\n",
+    "pandarallel.initialize()\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "eb99774f-9bcc-454d-b5e5-a8470223d6ca",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from rdkit import Chem\n",
+    "def make_canonical(smi):\n",
+    "    try:\n",
+    "        return Chem.MolToSmiles(Chem.MolFromSmiles(smi))\n",
+    "    except:\n",
+    "        return smi"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "id": "4d44bd8e-f2e1-44b4-aea7-40b4437baf44",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "df['smiles_can'] = df['smiles'].parallel_apply(make_canonical)"
    ]
   },
   {
@@ -1449,7 +1494,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 26,
+   "execution_count": 16,
    "id": "c6c64066-4032-4247-a8b9-00388176cc7b",
    "metadata": {},
    "outputs": [],
@@ -1459,17 +1504,19 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 9,
+   "execution_count": 18,
    "id": "469cf0dd-7b87-4245-973c-2a445e1fcca9",
    "metadata": {},
    "outputs": [
     {
      "data": {
       "text/plain": [
-       "Index(['seq', 'smiles', 'affinity_uM', 'neg_log10_affinity_M', 'affinity'], dtype='object')"
+       "Index(['seq', 'smiles', 'affinity_uM', 'neg_log10_affinity_M', 'affinity',\n",
+       "       'smiles_can'],\n",
+       "      dtype='object')"
       ]
      },
-     "execution_count": 9,
+     "execution_count": 18,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -1480,7 +1527,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 8,
+   "execution_count": 19,
    "id": "d91c0d91-474c-4ab2-9a5e-3b7861f7a832",
    "metadata": {},
    "outputs": [
@@ -1506,23 +1553,23 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 16,
+   "execution_count": 6,
    "id": "9ca8df46-15d3-40f9-b304-dd6e5597be5e",
    "metadata": {},
    "outputs": [
     {
      "data": {
       "text/plain": [
-       "2938"
+       "0.17005836848632214"
       ]
      },
-     "execution_count": 16,
+     "execution_count": 6,
      "metadata": {},
      "output_type": "execute_result"
     }
    ],
    "source": [
-    "(df['neg_log10_affinity_M']<2).sum()"
+    "(df['neg_log10_affinity_M']<5).sum()/len(df)"
    ]
   },
   {

data/all.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d22b80e9d2f1c3db65b792395b10f84dfd367dbe908cfd798aea5a10f1c62cb9
-size 204641362
+oid sha256:4864292b2aa4e63ffdc28ebdc7baea53a3a396d6e66ccd9927a04885586d160e
+size 228485896