SMARTS Sets

---

Overview

SMARTS queries are used for substructure searches and identifying functional groups. This module offers "sets" of common substructures and functional groups, which can be used to label molecules.

The SmartsSet class manages the search and labeling process for input molecules.

---

Available Sets

Two default SMARTS sets are available:

• ChemblAlerts (csv)
• CdkFunctionalGroups (csv)

The ChemblAlerts set includes several popular subsets:

• Glaxo
• Dundee
• BMS
• PAINS
• MLSMR

---

Basic Usage

Accessing a Set's Data

You can access the original raw data (from a CSV file) and a column for SMARTS Molecule objects using the smarts_data and smarts_dict class properties.

For instance, using ChemblAlerts:

from simmate.toolkit.smarts_sets import ChemblAlerts

# option 1 (as pandas.DataFrame)
data = ChemblAlerts.smarts_data

# option 2 (as dict)
data = ChemblAlerts.smarts_dict

Counting Matches

To obtain the exact count of each functional group, use get_counts:

from simmate.toolkit.smarts_sets import ChemblAlerts

matches = ChemblAlerts.get_counts(
    molecule,
    include_misses=False,
)

Listing Matches

If you only need a list of the SMARTS that matched, rather than exact counts, use get_matches:

from simmate.toolkit.smarts_sets import ChemblAlerts

matches = ChemblAlerts.get_matches(molecule)

---

Parallelization

Warning

The parallelization method for this class is not yet available. Please contact us if you require this feature.

---

Adding a New Set

To add a new SMARTS set, you need to:

1. Create a csv file with name and smarts_str columns (you can add more columns if needed)
2. Inherit from the SmartsSet base class
3. Specify the location of the CSV file with the source_file attribute

The SmartsSet will then manage the behavior of get_counts and other features.

For example:

from simmate.toolkit.smarts_sets.base import SmartsSet


class CustomAlerts(SmartsSet):
    source_file = "path/to/custom_alerts.csv"

---