Full Documentation

Note

Python 3 is required to run proteinTools. Any version of python 2 is currently not supported.

Creating Proteins

To create a protein structure, you can use the Protein(<Identifier>, <protein-ID> <species>) initializer, where species is optional (human by default) and protein-ID is also optional (PDB by default). Identifiers can be ChEMBL IDs, UniprotKB/Uniprot IDs, PDB IDs, or HGNC/Genecard IDs.

The protein file can than be downloaded to a directory (default the user’s current directory) with the .download method. Upon downloading the protein file (.pdb/.mmCIF from RCSB in the case of a PDB ID, or an Alphafold-generated structure otherwise), the protein will be populated with chains, residues, atoms, and other identifiers from the file.

from proteinTools import PT as p
myprot = p.Protein('2D3Z', 'e coli')
myprot.download('/path/to/directory')

newprot = p.Protein('P02786', 'Uniprot', 'human')
myprot.download('/path/to/directory')

newprot = p.Protein('L2HGDH', 'HGNC')
myprot.download('/path/to/directory')

Protein Properties

The protein’s ChEMBL ID, Ensembl ID, Uniprot ID, PDB ID, and Gene ID can all be accessed with properties of the same name. For PDB IDs, proteinTools returns a dataframe containing the PDB ID(s), minimum resolution of the PDB file in Angstroms, and the total number of unique ligands present in the file. The protein file does not need to be downloaded to access these values.

Uniprot = myprot.Uniprot
Gene = myprot.Gene
ChEMBL = myprot.ChEMBL
PDB = myprot.PDB
Ensembl = myprot.Ensembl

Protein residues can be queried by indexing as one would a list.

myprot[100:350]
myprot[100]

Residues and atoms can also be accessed via the .atoms and .residues method, which return the corrosponding atom and residue object.

myprot.residues('A_144')
myprot.atom(6572)

The protein’s FASTA sequence can be easily accessed by the FASTA property, which generates a FASTA sequence directly from the file.

myprot.FASTA

Protein-ligand interactions can be accessed with the ligand_interactions property, returning a dictionary containing BindingDB ligands and their activities, ChEMBL ligands and their activities (all in uM) and the ID of STITCH ligands and proteins. The protein does not need to be downloaded to access these values.

myprot.ligand_interactions

Protein-protein interactions can be accessed with the protein_interactions property, returning a dictionary containing stringDB protein keys with a list containing the maximum interaction score for the PPI and mean interaction score for the PPI. The protein does not need to be downloaded to access these values.

myprot.protein_interactions

The total amount of residues in the protein is obtainable simply by using the len() magic method.

protein_length = len(myprot)

A list of every atom in the protein and their properties can be created with the .to_csv(<destination>) method, where the default destination is the user’s current directory.

myprot.to_csv('/path/to/directory')

Every line of the file is present in the .data method.

protein_lines = myprot.data

the .ligands property returns a two-row dataframe, with the first row containing the primary ligand, all unique cofactors, and all ions present in the protein structure, where the second row contains the corrosponding ligand objects populated directly from the file for each category.

unique_ligands = myprot.ligands

Ligands can be queried from the .ligand_list attribute, which returns a list of every unique ligand position of every ligand present in the structural file (outside of water moelecules). Ligand properties are described below.

for ligand in myprot.ligand_list:
      ligands.append(ligand)

Note

CIF (mmCIF) files are currently supported by proteinTools, but certain functionalities (stripping ligand sites and secondary structure) are not available.

Residue Properties

Residue amino acids (AA), chain, atoms, index, and name can be accessed by properties of the same title.

myprot[1].name
myprot.residues('A433')['name']
myprot[5].AA
myprot[8].chain
myprot[2].atoms
residues = myprot[1:100]

The center of mass of each residue can be calculated with the .center property, which returns a list of the x, y, and z coordinate of the residue center.

residue_center = myprot[1].center

If the protein has a PDB ID format, the secondary structure of each residue can also be obtained with the structure property (between HELIX, SHEET, and UNSTRUCTURED).

residue_structure = myprot[160].structure

Atom Properties

The x, y, and z coordinate of atoms, as well as their mass, center element, line (line data from protein file), and atomic radius can be accessed by properties of the same title.

residue, elements, positions = myprot.residue('B123'), [], []
for atom in residue.atoms:
     elements.append(atom.element)
     atom_volume = 4 / 3 * math.pi * pow(atom.radius, 3)
     positions.append(atom.center)

The parent residue for atoms constructed via a protein file can be accessed with the .parent_residues property.

for atoms in residue.atoms:
   protein_chain = atoms.parent_residue.chain

Ligand Properties

If the protein is a PDB file containing ligands (that are not water molecules), they will automatically be added to the .ligands protein attribute. The ligand ID as present in the PDB file can be accessed with the ID attribute, and atoms of the atom class can be accessed with the atoms attribute.

The center of mass and the radius of gyration of each ligand can be calculated via their respective properties.

ligand.center
ligand.radius

The ligand file can be downloaded by the .download('/path/to/file') method, which defaults to the user’s current directory and saves the ligand in .sdf format.

ligand = protein.ligand_list[3]
ligand.download()
for ligand in protein.ligand_list:
    print(ligand.ID)
    print(ligand.center)

Ligand objects can also be created by specifying the path of the existing file when instantializing the ligand. Acceptable file formats are .pdb, .sdf, or .pdbqt.

ligand = protein.ligand(file_path = '/path/to/file')
center = ligand.center

Ligand files can also be instantiated separate of the protein. Simply generate the ligand with the ligand ID, and use the .download method with the path to download (defaults to the current user directory). InChiKeys, PDB IDs, and SMILEs sequences are accepted.

lig = p.ligand('C1=CC=C2C(=C1)C=CC=C2CCC(CO)N3C=C(N=C3)C(=O)N')
lig.download('/path/to/directory')

Warning

The above functionality depends on the chemical structure package Openbabel, which can be installed via pip install openbabel

Autodock Vina Integration

The protein file can be converted to .pdbqt format (a specialized Vina protein file format for docking) using the .pdbqt(‘/path/to/directory’) method, where the path represents the location of the downloaded pdb file, and defaults to the current user directory.

protein.pdbqt()

A recreation of the protein file stripped of ligand molecules can be created using the .strip_ligands method, where the destination is the location of the file, defaulting to the current user directory.

myprot.strip_ligands()

A recreation of the protein file stripped of only the primary ligand molecules (not cofactors and ions) can be created using the .strip_primary method, where the destination is the location of the file, defaulting to the current user directory.

myprot.strip_primary()