schrodinger.application.transforms.autosar.transforms module

AutoSAR is a workflow to explore combinations of RGroups at multiple positions while ensuring that the properties of the resulting molecule are within certain cutoffs. It is assumed that properties are additive.

Suppose we have a molecule where we have enumerated RGroups separately for R1, R2, and R3. AutoSAR will explore all possible combinations of RGroups at R1, R2, and R3 that satisfy the cutoffs. We estimate the entire molecule’s properties by summing the individual contributions of each RGroup.

class schrodinger.application.transforms.autosar.transforms.Combination(rgroups: ~typing.Tuple[~schrodinger.application.transforms.autosar.util.RGroup, ...] = <factory>)

Bases: object

Represent a combination of different RGroups on the same molecule and their aggregated properties.

When accessing a property using the index operator (e.g., combination[prop]), it will return the sum of that property from all RGroups in the combination.

Parameters:

rgroups – A tuple of RGroups that make up the combination

rgroups: Tuple[RGroup, ...]
toCSVRow(headers: list[str]) → str

Convert the Combination to a CSV row string based on the given headers.

Parameters:

headers – A list of headers to include in the CSV row.

Returns:

A CSV string representing the combination.

append(rgroup: RGroup) → Combination

Return a new Combination with the given RGroup appended.

NOTE: this does not modify the current combination.

__init__(rgroups: ~typing.Tuple[~schrodinger.application.transforms.autosar.util.RGroup, ...] = <factory>) → None
class schrodinger.application.transforms.autosar.transforms.ExpandCombination(rgroups: list[RGroup], cutoffs: dict[str, Cutoff], algorithm: Algorithm = Algorithm.SIMPLE)

Bases: PTransform

Expand a Combination by trying to add one more RGroup. All resulting Combinations will have their aggregated properties be less than the maximum value specified in the cutoffs.

NOTE: It is not guaranteed the resulting Combination’s aggregated properties will be larger than the minimum value specified in the cutoffs.

class Algorithm

Bases: Enum

SIMPLE = 'simple'
RANGE_SORT = 'range-sort'
__init__(rgroups: list[RGroup], cutoffs: dict[str, Cutoff], algorithm: Algorithm = Algorithm.SIMPLE)
Parameters:

  • rgroups – A list of RGroups to expand the combination with

  • cutoffs – A dictionary mapping property names to the min / max, inclusive cutoffs

  • algorithm – The algorithm to use to expand the combination

expand(combinations)
class schrodinger.application.transforms.autosar.transforms.EnumerateCombination(rgroups: list[RGroup])

Bases: PTransform

Enumerate all Combinations of RGroups with the input Combination

__init__(rgroups: list[RGroup])
Parameters:

rgroups – A list of RGroups to expand the combination with

expand(combinations)
class schrodinger.application.transforms.autosar.transforms.FilterCombination(cutoffs: dict[str, Cutoff])

Bases: PTransform

Filter out Combinations that are not within the min / max inclusive Cutoffs.

__init__(cutoffs: dict[str, Cutoff])
Parameters:

cutoffs – A dictionary mapping property names to the min / max inclusive cutoff.

expand(combinations)
schrodinger.application.transforms.autosar.transforms.is_within_cutoffs(combination: Combination, cutoffs: dict[str, Cutoff]) → bool

Is the given Combination within the Cutoff for each property

class schrodinger.application.transforms.autosar.transforms.ComputeNormalizedPropertiesSettings(reference_smiles: str, smiles_key: str, reference_properties: dict[str, float] = <factory>, properties_to_calculate: list[str] = <factory>, properties_to_extract: list[str] = <factory>, id_key: str = 'ID', atom_map_num: int = 1)

Bases: object

Parameters:

  • reference_smiles – The SMILES of the reference molecule

  • smiles_key – The key in the input dictionary that contains the SMILES

  • reference_properties – A dictionary of statically defined reference properties

  • properties_to_calculate – A list of properties to calculate for the input molecule and the reference molecule

  • properties_to_extract – A list of static properties to extract from the input dictionary and the reference_properties dict

  • id_key – The key to use for the ID of the output dictionary

  • atom_map_num – The atom map number to use for the RGroup attachment point in the output RGoup SMILES

reference_smiles: str
smiles_key: str
reference_properties: dict[str, float]
properties_to_calculate: list[str]
properties_to_extract: list[str]
id_key: str = 'ID'
atom_map_num: int = 1
validate()
__init__(reference_smiles: str, smiles_key: str, reference_properties: dict[str, float] = <factory>, properties_to_calculate: list[str] = <factory>, properties_to_extract: list[str] = <factory>, id_key: str = 'ID', atom_map_num: int = 1) → None
class schrodinger.application.transforms.autosar.transforms.ComputeNormalizedProperties(settings: ComputeNormalizedPropertiesSettings)

Bases: PTransform

Compute the normalized properties of a molecule by subtracting the reference properties from the molecule’s properties.

__init__(settings: ComputeNormalizedPropertiesSettings)
expand(rows)
schrodinger.application.transforms.autosar.transforms.calculate_properties(mol, property_keys: list[str]) → dict[str, float]
schrodinger.application.transforms.autosar.transforms.extract_properties(properties: dict[str, float], property_keys: list[str]) → dict[str, float]
schrodinger.application.transforms.autosar.transforms.get_rgroup_smiles(reference_smiles: str, target_smiles: str, atom_map_num: int = 1) → str

Determine the RGroup SMILES assuming the reference_smiles is a core of the target_smiles. The attachment point on the RGroup will have the atom map number set to atom_map_num.

Parameters:

  • reference_smiles – The SMILES of the core molecule

  • target_smiles – The SMILES of the target molecule

  • atom_map_num – The atom map number to set on the RGroup attachment point