Skip to content

Fingerprints

molfp.fingerprints

All 13 molecular fingerprint functions.

Each function accepts
  • A single SMILES string or RDKit Mol → returns list[int]
  • A list of SMILES/Mols → returns numpy ndarray (n_samples, n_bits)
  • A pandas Series → returns numpy ndarray (n_samples, n_bits)
  • A pandas DataFrame + column name → returns numpy ndarray (n_samples, n_bits)

maccs(input_data, *, column=None, n_jobs=1, sparse=False)

MACCS Keys fingerprint (166 bits).

ecfp4(input_data, *, n_bits=1024, collision_free=False, column=None, n_jobs=1, sparse=False)

ECFP4 / Morgan radius=2 fingerprint. Use collision_free=True for multi-hash version.

fcfp4(input_data, *, n_bits=1024, collision_free=False, column=None, n_jobs=1, sparse=False)

FCFP4 / Feature Morgan radius=2 fingerprint. Uses pharmacophoric atom types instead of atomic invariants.

pubchem(input_data, *, column=None, n_jobs=1, sparse=False)

PubChem/CACTVS fingerprint (881 bits).

substructure(input_data, *, column=None, n_jobs=1, sparse=False)

SubstructureFP (307 bits) — Inte:Ligand SMARTS patterns.

substructure_count(input_data, *, column=None, n_jobs=1, sparse=False)

Substructure Count (307 features) — match counts.

klekota_roth(input_data, *, column=None, n_jobs=1, sparse=False)

Klekota-Roth fingerprint (4860 bits).

klekota_roth_count(input_data, *, column=None, n_jobs=1, sparse=False)

Klekota-Roth Count (4860 features) — match counts.

estate(input_data, *, column=None, n_jobs=1, sparse=False)

E-State fingerprint (79 bits).

atompairs2d(input_data, *, n_bits=780, collision_free=False, column=None, n_jobs=1, sparse=False)

AtomPairs2D fingerprint (780 bits, hashed). Use collision_free=True for multi-hash.

atompairs2d_count(input_data, *, n_bits=780, collision_free=False, column=None, n_jobs=1, sparse=False)

AtomPairs2D Count (780 features, hashed counts). Use collision_free=True for multi-hash.

rdk(input_data, *, n_bits=1024, collision_free=False, column=None, n_jobs=1, sparse=False)

RDKit path-based fingerprint (1024 bits). Use collision_free=True for multi-hash.

rdk_ext(input_data, *, n_bits=1024, collision_free=False, column=None, n_jobs=1, sparse=False)

RDKit Extended fingerprint (1024 bits). Use collision_free=True for multi-hash.

rdk_graphonly(input_data, *, n_bits=1024, collision_free=False, column=None, n_jobs=1, sparse=False)

RDKit GraphOnly fingerprint (1024 bits). Use collision_free=True for multi-hash.

lingo(input_data, *, n_bits=1024, column=None, n_jobs=1, sparse=False)

Lingo fingerprint (hashed SMILES 4-grams).

shortest_path(input_data, *, n_bits=1024, column=None, n_jobs=1, sparse=False)

Shortest Path fingerprint — encodes atom-pair features along shortest paths.

hybridization(input_data, *, n_bits=1024, column=None, n_jobs=1, sparse=False)

Hybridization fingerprint — path-based with atom hybridization invariants.

signature(input_data, *, n_bits=1024, column=None, n_jobs=1, sparse=False)

Signature fingerprint — atom-centered subgraph descriptors (Faulon signatures).

atompairs3d(input_data, *, n_bits=780, collision_free=False, column=None, n_jobs=1, sparse=False)

3D Atom Pairs fingerprint (780 bits). Use collision_free=True for multi-hash.

atompairs3d_count(input_data, *, n_bits=780, collision_free=False, column=None, n_jobs=1, sparse=False)

3D Atom Pairs Count (780 features). Use collision_free=True for multi-hash.

usr(input_data, *, column=None, n_jobs=1, sparse=False)

Ultrafast Shape Recognition (12 floats) — molecular shape descriptor.

usrcat(input_data, *, column=None, n_jobs=1, sparse=False)

USRCAT (60 floats) — USR + pharmacophoric atom categories.

autocorr3d(input_data, *, column=None, n_jobs=1, sparse=False)

3D Autocorrelation (80 floats) — spatial correlation of atomic properties.

rdf(input_data, *, column=None, n_jobs=1, sparse=False)

Radial Distribution Function (210 floats) — atom-pair distance histogram.

morse(input_data, *, column=None, n_jobs=1, sparse=False)

MORSE (224 floats) — molecule representation based on electron diffraction.

whim(input_data, *, column=None, n_jobs=1, sparse=False)

WHIM (114 floats) — weighted holistic invariant molecular descriptors.

getaway(input_data, *, column=None, n_jobs=1, sparse=False)

GETAWAY (273 floats) — geometry, topology, and atom-weights assembly.

available()

Return list of available fingerprint names.

get(name)

Get fingerprint function by name. Returns (fn, n_bits, description).