def compute_morgan_fingerprints(graph: nx.Graph, fingerprint_length: int, fingerprint_radius: int):
"""Get Morgan Fingerprint of a specific SMILES string.
Adapted from: <https://github.com/google-research/google-research/blob/
dfac4178ccf521e8d6eae45f7b0a33a6a5b691ee/mol_dqn/chemgraph/dqn/deep_q_networks.py#L750>
Args:
graph (nx.Graph): The molecule as a networkx graph
fingerprint_length (int): Bit-length of fingerprint
fingerprint_radius (int): Radius used to compute fingerprint
Returns:
np.array. shape = [hparams.fingerprint_length]. The Morgan fingerprint.
"""
if graph is None: # No smiles string
return np.zeros((fingerprint_length,))
molecule = convert_nx_to_rdkit(graph)
if molecule is None: # Invalid smiles string
return np.zeros((fingerprint_length,))
# Compute the fingerprint
fingerprint = AllChem.GetMorganFingerprintAsBitVect(
molecule, fingerprint_radius, fingerprint_length)
arr = np.zeros((1,))
# ConvertToNumpyArray takes ~ 0.19 ms, while
# np.asarray takes ~ 4.69 ms
DataStructs.ConvertToNumpyArray(fingerprint, arr)
return arr
def _call(self, graph: nx.Graph) -> float:
# Compute all the rewards
mol = convert_nx_to_rdkit(graph)
mw = Descriptors.MolWt(mol)
pic50 = self.pic50_reward(graph)
sim = self.similarity_reward(graph)
# Normalize them
mw_score = 1.0 / ((1 + np.exp(
(self.mw_range[0] - mw) / self.mw_tol)) * (1 + np.exp(
(mw - self.mw_range[1]) / self.mw_tol)))
pic50_score = 1.0 / (1 + np.exp((8 - pic50) / self.pic50_tol))
# Sim is already between [0, 1]
# Combine them
return sim + mw_score + pic50_score
def _call(self, graph: nx.Graph) -> float:
mol = convert_nx_to_rdkit(graph)
return QED.qed(mol)
def _call(self, graph: nx.Graph) -> float:
mol = convert_nx_to_rdkit(graph)
return calculateScore(mol)
def _call(self, graph: nx.Graph) -> float:
mol = convert_nx_to_rdkit(graph)
return Crippen.MolLogP(mol)