def klifs_kinase_from_uniprot_id(uniprot_id: str) -> pd.DataFrame:
"""
Retrieve KLIFS kinase details about the kinase matching the given Uniprot ID.
Parameters
----------
uniprot_id: str
Uniprot identifier.
Returns
-------
kinase: pd.Series
KLIFS kinase details.
Raises
------
ValueError:
No KLIFS kinase found for UniProt ID.
ValueError:
Multiple KLIFS kinases found for UniProt ID.
"""
from opencadd.databases.klifs import setup_remote
remote = setup_remote()
kinase_ids = remote.kinases.all_kinases()["kinase.klifs_id"]
kinases = remote.kinases.by_kinase_klifs_id(list(kinase_ids))
kinases = kinases[kinases["kinase.uniprot"] == uniprot_id]
if len(kinases) == 0:
raise ValueError("No KLIFS kinase found for UniProt ID.")
elif len(kinases) > 1:
raise ValueError("Multiple KLIFS kinases found for UniProt ID.")
kinase = kinases.iloc[0]
return kinase
def from_structure_klifs_id(cls,
structure_klifs_id,
subpockets=None,
extension="pdb",
klifs_session=None):
"""
Get a KLIFS pocket (remotely by a structure KLIFS ID) that defines the KLIFS regions and
subpockets.
Parameters
----------
structure_klifs_id : int
Structure KLIFS ID.
subpockets : dict
Dictionary with the following keys and values:
"anchor_residue.klifs_id" : list of int
List of anchor residues (KLIFS residue IDs) whose centroid defines the subpocket
center.
"subpocket.name" : str
Subpocket name.
"subpocket.color" : str
Subpocket color.
extension : str
Structure protein data file format. Defaults to PDB format.
klifs_session : opencadd.databases.klifs.session.Session or None
Remote or local KLIFS session. If None, a remote session is initialized.
Returns
-------
opencadd.structure.pocket.PocketKlifs
KLIFS pocket object.
"""
# Use existing KLIFS session or set up remote session
if not klifs_session:
klifs_session = setup_remote()
# Get pocket and coordinates for a structure (by a structure KLIFS ID)
if klifs_session._client:
pocket_residues = klifs_session.pockets.by_structure_klifs_id(
structure_klifs_id)
else:
pocket_residues = klifs_session.pockets.by_structure_klifs_id(
structure_klifs_id, extension=extension)
text = klifs_session.coordinates.to_text(structure_klifs_id,
entity="complex",
extension=extension)
pocket = cls.from_text(
text,
extension,
pocket_residues["residue.id"].to_list(),
pocket_residues["residue.klifs_id"].to_list(),
structure_klifs_id,
)
pocket = pocket.add_klifs_regions(pocket, pocket_residues)
pocket = pocket.add_klifs_subpockets(pocket, pocket_residues,
subpockets)
return pocket
def _to_kinase_annotation(distance_matrix, outputfile):
"""
Save kinase annotations to file used for FigTree.
Parameters
----------
distance_matrix : pandas.DataFrame
Distance matrix on which clustering is based.
outputfile : str or pathlib.Path
Path to kinase annotation file (CSV file) in FigTree format.
"""
outputfile = Path(outputfile)
logger.info(f"Writing resulting kinase annotation to {outputfile}")
# Get kinase names from matrix
kinase_names = distance_matrix.columns.to_list()
# Query KLIFS for kinase details
klifs_session = klifs.setup_remote()
kinases = klifs_session.kinases.by_kinase_name(kinase_names)
kinases = kinases[kinases["species.klifs"] == "Human"]
kinases = kinases[["kinase.klifs_name", "kinase.family", "kinase.group"]]
# Save to file
kinases.to_csv(outputfile, sep="\t", index=False)
def plot_number_of_kinases_per_kinase_group(structures, remote=None):
"""
Plot the number of kinases per kinase group.
Parameters
----------
structures : pandas.DataFrame
Structures DataFrame from opencadd.databases.klifs module.
remote : None or opencadd.databases.klifs.session.Session
Remote KLIFS session. If None, generate new remote session.
Returns
-------
matplotlib.pyplot.axis
Plot axis.
"""
kinase_ids = structures["kinase.klifs_id"].to_list()
# Get kinases by kinase KLIFS IDs
if remote is None:
remote = setup_remote()
kinases = remote.kinases.by_kinase_klifs_id(kinase_ids)
ax = (kinases.groupby("kinase.group").size().sort_values().plot(
kind="barh",
figsize=(4, 3),
title="Number of kinases per kinase group",
xlabel="Kinase group",
))
return ax
def from_structure_klifs_ids(cls, structure_klifs_ids, klifs_session=None, n_cores=1):
"""
Calculate fingerprints for one or more KLIFS structures (by structure KLIFS IDs).
Parameters
----------
structure_klifs_id : int
Input structure KLIFS ID (output fingerprints may contain less IDs because some
structures could not be encoded).
klifs_session : opencadd.databases.klifs.session.Session
Local or remote KLIFS session.
n_cores : int or None
Number of cores to be used for fingerprint generation as defined by the user.
Returns
-------
kissim.encoding.fingerprint_generator
Fingerprint generator object containing fingerprints.
"""
logger.info("GENERATE FINGERPRINTS")
logger.info(f"Number of input structures: {len(structure_klifs_ids)}")
start_time = datetime.datetime.now()
logger.info(f"Fingerprint generation started at: {start_time}")
# Set up KLIFS session if needed
if klifs_session is None:
klifs_session = setup_remote()
# Set number of cores to be used
n_cores = set_n_cores(n_cores)
# Initialize FingerprintGenerator object
fingerprint_generator = cls()
fingerprint_generator.structure_klifs_ids = structure_klifs_ids
fingerprint_generator.klifs_session = klifs_session
fingerprints_list = fingerprint_generator._get_fingerprint_list(n_cores)
fingerprint_generator.data = {
i.structure_klifs_id: i
for i in fingerprints_list
if i is not None # Removes emtpy fingerprints
}
fingerprint_generator.data_normalized = fingerprint_generator._normalize_fingerprints()
logger.info(f"Number of output fingerprints: {len(fingerprint_generator.data)}")
end_time = datetime.datetime.now()
logger.info(f"Runtime: {end_time - start_time}")
return fingerprint_generator
def _get_klifs_residue_colors(remote=None):
"""
Get KLIFS residue colors from example structure KLIFS ID (12347).
Parameters
----------
remote : None or opencadd.databases.klifs.session.Session
Remote KLIFS session. If None, generate new remote session.
Returns
-------
list of str
KLIFS residue colors (matplotlib color names).
"""
if remote is None:
remote = setup_remote()
klifs_colors = remote.pockets.by_structure_klifs_id(
12347)["residue.klifs_color"]
return klifs_colors.to_list()
def _setup_klifs_session(local_klifs_download_path=None):
"""
Set up KLIFS session.
Parameters
----------
local_klifs_download_path : str or None
If path to local KLIFS download is given, set up local KLIFS session.
If None is given, set up remote KLIFS session.
Returns
-------
klifs_session : opencadd.databases.klifs.session.Session
Local or remote KLIFS session.
"""
if local_klifs_download_path:
klifs_session = setup_local(local_klifs_download_path)
else:
klifs_session = setup_remote()
return klifs_session
def from_structure_klifs_id(cls, structure_klifs_id, klifs_session=None):
"""
Get KLIFS data from structure KLIFS ID.
Parameters
----------
structure_klifs_id : int
KLIFS structure ID.
klifs_session : opencadd.databases.klifs.session.Session
Local or remote KLIFS session.
Returns
-------
kissim.io.KlifsToKissimData
KLIFS data.
"""
data = cls()
data.structure_klifs_id = structure_klifs_id
# If no KLIFS session is given, set up remote KLIFS session
if klifs_session is None:
klifs_session = setup_remote()
data.klifs_session = klifs_session
# Structure KLIFS ID exists
if not data._structure_klifs_id_exists():
return None
# In case of a local KLIFS session, test if complex and pocket structural files exist
if data.klifs_session._database is not None:
if not data._local_session_files_exist():
return None
data.text, data.extension = data._get_text_and_extension()
data.residue_ids, data.residue_ixs = data._get_pocket_residue_ids_and_ixs()
data.kinase_name = data._get_kinase_name()
return data
def from_structure_klifs_id(cls, structure_klifs_id, subpockets=None):
"""
Get a KLIFS pocket (remotely by a structure KLIFS ID) that defines the KLIFS regions and
subpockets.
Parameters
----------
structure_klifs_id : int
Structure KLIFS ID.
subpockets : pandas.DataFrame
Subpockets (row) with the following details (columns):
"anchor_residue.klifs_id" : list of int
List of anchor residues (KLIFS residue IDs) whose centroid defines the subpocket
center.
"subpocket.name" : str
Subpocket name.
"subpocket.color" : str
Subpocket color.
Returns
-------
opencadd.structure.pocket.KlifsPocket
KLIFS pocket object.
"""
# Set up remote KLIFS session
remote = setup_remote()
# Get pocket and coordinates for a structure (by a structure KLIFS ID)
pocket = remote.pockets.by_structure_klifs_id(structure_klifs_id)
filepath = remote.coordinates.to_pdb(structure_klifs_id,
".",
entity="complex")
pocket_3d = cls.from_file(
filepath,
pocket["residue.id"].to_list(),
"example kinase",
pocket["residue.klifs_id"].to_list(),
)
# Add regions
for (region, color), group in pocket.groupby(
["residue.klifs_region_id", "residue.klifs_color"]):
pocket_3d.add_region(
region,
group["residue.id"].to_list(),
color,
group["residue.klifs_region_id"].to_list(),
)
# Map residue KLIFS IDs > residue ID
if subpockets is not None:
subpockets["anchor_residue.ids"] = subpockets[
"anchor_residue.klifs_ids"].apply(lambda x: pocket[pocket[
"residue.klifs_id"].isin(x)]["residue.id"].to_list())
# Add subpockets
for _, subpocket in subpockets.iterrows():
pocket_3d.add_subpocket(
subpocket["subpocket.name"],
subpocket["anchor_residue.ids"],
subpocket["subpocket.color"],
subpocket["anchor_residue.klifs_ids"],
)
return pocket_3d
import pytest
import numpy as np
import pandas as pd
from opencadd.databases.klifs import setup_local, setup_remote
from kissim.utils import enter_temp_directory
from kissim.encoding import Fingerprint, FingerprintNormalized, FingerprintGenerator
from kissim.schema import (
FEATURE_NAMES_PHYSICOCHEMICAL,
FEATURE_NAMES_PHYSICOCHEMICAL_DICT,
FEATURE_NAMES_DISTANCES_AND_MOMENTS,
)
PATH_TEST_DATA = Path(__name__).parent / "kissim" / "tests" / "data"
REMOTE = setup_remote()
LOCAL = setup_local(PATH_TEST_DATA / "KLIFS_download")
class TestFingerprintGenerator:
"""
Test common functionalities in the PocketBioPython and PocketDataFrame classes.
"""
@pytest.mark.parametrize(
"structure_klifs_ids, klifs_session, n_cores, fingerprints_values_array_sum",
[
([110, 118], REMOTE, 1, 10152.4256),
([110, 118], REMOTE, 2, 10152.4256),
([110, 118], LOCAL, 1, 10152.4256),
([110, 118], LOCAL, 2, 10152.4256),
([110, 118], None, None, 10152.4256),
Draw,
Descriptors,
Lipinski,
PandasTools,
rdFingerprintGenerator,
QED,
)
from rdkit.Chem.Draw import IPythonConsole
from rdkit.Chem.MolStandardize import rdMolStandardize
from rdkit.Chem.PropertyMol import PropertyMol
from rdkit.ML.Cluster import Butina
import seaborn as sns
from opencadd.databases.klifs import setup_remote
KLIFS_SESSION = setup_remote()
RDLogger.DisableLog("rdApp.*")
SUBPOCKET_COLORS = {
"AP": "purple",
"FP": "forestgreen",
"SE": "c",
"GA": "tab:orange",
"B1": "tab:blue",
"B2": "darkslateblue",
"X": "grey",
}
def read_fragment_library(path_to_lib):
def from_structure_klifs_id(
cls, structure_klifs_id, subpockets=None, extension="pdb", klifs_session=None
):
"""
Get a KLIFS pocket (remotely by a structure KLIFS ID) that defines the KLIFS regions and
subpockets.
Parameters
----------
structure_klifs_id : int
Structure KLIFS ID.
subpockets : dict
Dictionary with the following keys and values:
"anchor_residue.klifs_id" : list of int
List of anchor residues (KLIFS residue IDs) whose centroid defines the subpocket
center.
"subpocket.name" : str
Subpocket name.
"subpocket.color" : str
Subpocket color.
extension : str
Structure protein data file format. Defaults to PDB format.
klifs_session : opencadd.databases.klifs.session.Session or None
Remote or local KLIFS session. If None, a remote session is initialized.
Returns
-------
opencadd.structure.pocket.KlifsPocket
KLIFS pocket object.
"""
# Use existing KLIFS session or set up remote session
if not klifs_session:
klifs_session = setup_remote()
# Get pocket and coordinates for a structure (by a structure KLIFS ID)
if klifs_session._client:
pocket = klifs_session.pockets.by_structure_klifs_id(structure_klifs_id)
else:
pocket = klifs_session.pockets.by_structure_klifs_id(
structure_klifs_id, extension=extension
)
text = klifs_session.coordinates.to_text(
structure_klifs_id, entity="complex", extension=extension
)
pocket_3d = cls.from_text(
text,
extension,
pocket["residue.id"].to_list(),
pocket["residue.klifs_id"].to_list(),
structure_klifs_id,
)
# Add regions
for (region, color), group in pocket.groupby(
["residue.klifs_region", "residue.klifs_color"]
):
pocket_3d.add_region(
name=region,
residue_ixs=group["residue.klifs_id"].to_list(),
color=color,
)
# Map residue KLIFS IDs > residue ID
if subpockets is not None:
subpockets = pd.DataFrame(subpockets)
subpockets["anchor_residue.ids"] = subpockets["anchor_residue.klifs_ids"].apply(
lambda x: pocket[pocket["residue.klifs_id"].isin(x)]["residue.id"].to_list()
)
# Add subpockets
for _, subpocket in subpockets.iterrows():
pocket_3d.add_subpocket(
name=subpocket["subpocket.name"],
anchor_residue_ixs=subpocket["anchor_residue.klifs_ids"],
color=subpocket["subpocket.color"],
)
return pocket_3d