Fingerprints and Descriptors

class Smi2Fp(radius=3, fpSize=2048)[source]

Calculate Morgan fingerprints from SMILES strings

get_np(smiles)[source]

Convert a SMILES string to a numpy array with Morgan fingerprint bits.

Parameters:: smiles – SMILES string
Returns:: numpy array with Morgan fingerprint bits

get_np_counts(smiles)[source]

Convert a SMILES string to a numpy array with Morgan fingerprint counts.

Parameters:: smiles – SMILES string
Returns:: numpy array with Morgan fingerprint counts

get_fp(smiles)[source]

Convert a SMILES string to a Morgan fingerprint.

Parameters:: smiles – SMILES string
Returns:: Morgan fingerprint

get_count_fp(smiles)[source]

Convert a SMILES string to a Morgan count fingerprint.

Parameters:: smiles – SMILES string
Returns:: Morgan count fingerprint

mol2morgan_fp(mol, radius=2, nBits=2048)[source]

Convert an RDKit molecule to a Morgan fingerprint To avoid the rdkit deprecated warning, do this from rdkit import rdBase with rdBase.BlockLogs():

uru.smi2numpy_fp(“CCC”)

Parameters:

mol (Mol) – RDKit molecule
radius (int) – fingerprint radius
nBits (int) – number of fingerprint bits

Return type:

ExplicitBitVect

Returns:

RDKit Morgan fingerprint

smi2morgan_fp(smi, radius=2, nBits=2048)[source]

Convert a SMILES to a Morgan fingerprint To avoid the rdkit deprecated warning, do this from rdkit import rdBase with rdBase.BlockLogs():

uru.smi2numpy_fp(“CCC”)

Parameters:

smi (str) – SMILES
radius (int) – fingerprint radius
nBits (int) – number of fingerprint bits

Return type:

Optional[ExplicitBitVect]

Returns:

RDKit Morgan fingerprint

mol2numpy_fp(mol, radius=2, n_bits=2048)[source]

Convert an RDKit molecule to a numpy array with Morgan fingerprint bits Borrowed from https://iwatobipen.wordpress.com/2019/02/08/convert-fingerprint-to-numpy-array-and-conver-numpy-array-to-fingerprint-rdkit-memorandum/

Parameters:

mol (Mol) – RDKit molecule
radius (int) – fingerprint radius
n_bits (int) – number of fingerprint bits

Return type:

ndarray

Returns:

numpy array with RDKit fingerprint bits

smi2numpy_fp(smi, radius=2, nBits=2048)[source]

Convert a SMILES to a numpy array with Morgan fingerprint bits

Parameters:

smi (str) – SMILES string
radius (int) – fingerprint radius
nBits (int) – number of fingerprint bits

Return type:

ndarray

Returns:

numpy array with RDKit fingerprint bits

class RDKitDescriptors(desc_names=None, hide_progress=False, skip_fragments=False)[source]

Calculate RDKit descriptors for molecules or SMILES.

Provide methods to compute descriptor vectors for a single molecule or SMILES, and to produce pandas DataFrames for lists of molecules or SMILES.

Attributes

desc_names: Sorted list of descriptor names that will be calculated.
hide_progress: Whether to hide progress bars when processing lists.

Initialize descriptor calculator.

type desc_names:
param desc_names:: Optional list of descriptor names to use. If not provided, the full RDKit descriptor list is used.
type hide_progress:: bool
param hide_progress:: If true, progress bars are disabled when processing lists.
type skip_fragments:
param skip_fragments:: If true, descriptors whose names contain “fr_” are excluded.
return:: None

update_descriptors(index_list)[source]

Update the descriptor names to only include those at the specified indices.

Parameters:: index_list (List[int]) – List of indices to keep
Return type:: None
Returns:: None

calc_mol(mol)[source]

Calculate descriptors for an RDKit molecule.

Parameters:: mol (Mol) – RDKit molecule
Return type:: ndarray
Returns:: A numpy array with descriptor values

calc_smiles(smiles)[source]

Calculate descriptors for a SMILES string.

Parameters:: smiles (str) – SMILES string
Return type:: ndarray
Returns:: A numpy array with descriptor values

pandas_smiles(smiles_list)[source]

Calculate descriptors for a list of SMILES and return a DataFrame.

Parameters:: smiles_list (List[str]) – List of SMILES strings
Return type:: DataFrame
Returns:: DataFrame where each row corresponds to a SMILES and columns are descriptors

pandas_mols(mol_list)[source]

Calculate descriptors for a list of RDKit molecules and return a DataFrame.

Parameters:: mol_list (List[Mol]) – List of RDKit molecule objects
Return type:: DataFrame
Returns:: DataFrame where each row corresponds to a molecule and columns are descriptors

clean_descriptors(desc_in)[source]

Remove descriptor columns that contain any NaN or infinite values.

Parameters:: desc_in (ndarray) – Input descriptor array
Return type:: tuple[Union[ndarray, List[int]]]
Returns:: Tuple containing the cleaned descriptor array (only columns with all finite values) and a list of kept column indices

scale_descriptors(desc_in)[source]

Scale descriptor DataFrame using StandardScaler.

Parameters:: desc_in (DataFrame) – Input descriptor DataFrame
Return type:: tuple[Any, StandardScaler]
Returns:: Tuple with the scaled descriptor array and the fitted StandardScaler

clean_and_scale_descriptors(desc_in)[source]

Clean and scale a descriptor DataFrame.

Parameters:: desc_in (DataFrame) – Input descriptor DataFrame
Return type:: tuple[Any, StandardScaler]
Returns:: Tuple containing the cleaned and scaled descriptor array and the fitted StandardScaler

class RDKitProperties[source]

Calculate RDKit properties

calc_mol(mol)[source]

Calculate properties for an RDKit molecule

Parameters:: mol (Mol) – RDKit molecule
Return type:: ndarray
Returns:: a numpy array with properties

calc_smiles(smi)[source]

Calculate properties for a SMILES string

Parameters:: smi (str) – SMILES string
Return type:: Optional[ndarray]
Returns:: a numpy array with properties

pandas_smiles(smi_list)[source]

Calculates properties for a list of SMILES strings and returns them as a pandas DataFrame.

Parameters:: smi_list (List[str]) – List of SMILES strings
Returns:: DataFrame with calculated properties. Each row corresponds to a SMILES string and each column to a property.
Return type:: pd.DataFrame

pandas_mols(mol_list)[source]

Calculates properties for a list of RDKit molecules and returns them as a pandas DataFrame.

Parameters:: mol_list (List[Mol]) – List of RDKit molecules
Returns:: DataFrame with calculated properties. Each row corresponds to a molecule and each column to a property.
Return type:: pd.DataFrame

class Ro5Calculator[source]

A class used to calculate Lipinski’s Rule of Five properties for a given molecule.

Attributes

namesList[str]: A list of names of the properties to be calculated.
functionsList[Callable[[Mol], float]]: A list of functions used to calculate the properties.

Methods

calc_mol(mol: Mol) -> np.ndarray: Calculates properties for a RDKit molecule.
calc_smiles(smi: str) -> Optional[np.ndarray]: Calculates properties for a SMILES string.
pandas_smiles(smiles_list: List[str]) -> pd.DataFrame: Calculates properties for a list of SMILES strings and returns them as a pandas DataFrame.
pandas_mols(mol_list: List[Mol]) -> pd.DataFrame: Calculates properties for a list of RDKit molecules and returns them as a pandas DataFrame.

Initialize the Ro5Calculator class.

type self:: Ro5Calculator
param self:: An instance of the Ro5Calculator class
type self:: Ro5Calculator
return:: None
rtype:: None

calc_mol(mol)[source]

Calculate properties for a RDKit molecule

Parameters:: mol (Mol) – RDKit molecule
Returns:: a numpy array with properties
Return type:: np.ndarray

calc_smiles(smi)[source]

Calculate properties for a SMILES string

Parameters:: smi (str) – SMILES string
Returns:: a numpy array with properties
Return type:: Optional[np.ndarray]

pandas_smiles(smiles_list)[source]

Calculates properties for a list of SMILES strings and returns them as a pandas DataFrame.

Parameters:: smiles_list (List[str]) – List of SMILES strings
Returns:: DataFrame with calculated properties. Each row corresponds to a SMILES string and each column to a property.
Return type:: pd.DataFrame

pandas_mols(mol_list)[source]

Calculates properties for a list of RDKit molecules and returns them as a pandas DataFrame.

Parameters:: mol_list (List[Mol]) – List of RDKit molecules
Returns:: DataFrame with calculated properties. Each row corresponds to a molecule and each column to a property.
Return type:: pd.DataFrame

compare_datasets(train_fp, test_fp)[source]

Compare two datasets of fingerprints and return the maximum Tanimoto similarity for each test fingerprint.

Parameters:

train_fp (list) – Training set fingerprints.
test_fp (list) – Test set fingerprints.

Return type:

list

Returns:

List of maximum similarity values for each test fingerprint.