feat: Add calculating Tanimoto similarity #117

app/modules/cdkmodules.py

@@ -237,3 +237,37 @@ def getCDKDescriptors(smiles: str):
             float("{:.2f}".format(float(str(FractionalCSP3Descriptor)))),
             int(NumRings),
         )
+
+
+def getTanimotoSimilarity(smiles1: str, smiles2: str):
+    """
+    Take two SMILES strings and calculate
+    Tanimoto similarity index using Pubchem
+    Fingerprints.
+    Args (str,str): SMILES strings.
+    Returns (float): Tanimoto similarity.
+    """
+    if any(char.isspace() for char in smiles1):
+        smiles1 = smiles1.replace(" ", "+")
+    if any(char.isspace() for char in smiles2):
+        smiles2 = smiles2.replace(" ", "+")
+
+    Tanimoto = JClass(cdk_base + ".similarity.Tanimoto")
+    SCOB = JClass(cdk_base + ".silent.SilentChemObjectBuilder")
+    SmilesParser = JClass(cdk_base + ".smiles.SmilesParser")(SCOB.getInstance())
+    PubchemFingerprinter = JClass(cdk_base + ".fingerprint.PubchemFingerprinter")(
+        SCOB.getInstance()
+    )
+
+    # parse molecules to get IAtomContainers
+    mol1 = SmilesParser.parseSmiles(smiles1)
+    mol2 = SmilesParser.parseSmiles(smiles2)
+
+    # Generate BitSets using PubChemFingerprinter
+    fingerprint1 = PubchemFingerprinter.getBitFingerprint(mol1).asBitSet()
+    fingerprint2 = PubchemFingerprinter.getBitFingerprint(mol2).asBitSet()
+
+    # Calculate Tanimoto similarity
+    Similarity = Tanimoto.calculate(fingerprint1, fingerprint2)
+
+    return "{:.5f}".format(float(str(Similarity)))

app/modules/rdkitmodules.py

@@ -1,5 +1,5 @@
 from chembl_structure_pipeline import standardizer
-from rdkit import Chem
+from rdkit import Chem, DataStructs
 from rdkit.Chem import AllChem, Descriptors, QED, Lipinski, rdMolDescriptors, rdmolops
 from app.modules.cdkmodules import getCDKSDGMol
 
@@ -112,3 +112,25 @@ def get3Dconformers(smiles, depict=True):
             return Chem.MolToMolBlock(mol)
     else:
         return "Error reading SMILES string, check again."
+
+
+def getTanimoto(smiles1, smiles2):
+    """
+    Take two SMILES strings and calculate
+    Tanimoto similarity index using Morgan
+    Fingerprints.
+    Args (str,str): SMILES strings.
+    Returns (float): Tanimoto similarity.
+    """
+    # create two example molecules
+    mol1 = checkSMILES(smiles1)
+    mol2 = checkSMILES(smiles2)
+
+    # generate Morgan fingerprints for each molecule
+    fp1 = AllChem.GetMorganFingerprintAsBitVect(mol1, 2, nBits=1024)
+    fp2 = AllChem.GetMorganFingerprintAsBitVect(mol2, 2, nBits=1024)
+
+    # calculate the Tanimoto similarity between the fingerprints
+    similarity = DataStructs.TanimotoSimilarity(fp1, fp2)
+
+    return similarity

app/routers/chem.py

@@ -8,9 +8,9 @@
 from fastapi.responses import Response, HTMLResponse
 from app.modules.npscorer import getNPScore
 from app.modules.classyfire import classify, result
-from app.modules.cdkmodules import getCDKSDGMol
+from app.modules.cdkmodules import getCDKSDGMol, getTanimotoSimilarity
 from app.modules.depict import getRDKitDepiction, getCDKDepiction
-from app.modules.rdkitmodules import get3Dconformers
+from app.modules.rdkitmodules import get3Dconformers, getTanimoto
 from app.modules.coconutdescriptors import getCOCONUTDescriptors
 import pandas as pd
 from fastapi.templating import Jinja2Templates
@@ -139,6 +139,17 @@ async def CDK2D_coordinates(smiles: str):
             return "Error reading SMILES string, check again."
 
 
+@router.get("/tanimoto")
+async def Tanimoto(smiles: str, toolkit: Optional[str] = "cdk"):
+    if smiles:
+        smiles1, smiles2 = smiles.split(",")
+        if toolkit == "rdkit":
+            Tanimoto = getTanimoto(smiles1, smiles2)
+        else:
+            Tanimoto = getTanimotoSimilarity(smiles1, smiles2)
+        return Tanimoto
+
+
 @router.get("/depict")
 async def depict_molecule(
     smiles: str,