API Reference#

The sections beneath contain all functions that can be used, ordered by module. The modules that are part of the C++ core are indicated by their section title. All functions have the same signature in the Python package and the C++ core. Only the properties of the Protein class are called differently. All API references are given in alphabetical order.

Algorithms - core#

Prospr contains a couple of built-in search algorithms that will find the most optimal conformation. These can all be imported directly from prospr without specifying a submodule, e.g.

from prospr import depth_first
depth_first(protein)
Finds the most optimal conformation using a depth-first algorithm.
Does not reset the Protein properties beforehand!
Parameters:
* protein - Protein: the Protein object to fold.
Returns:
* Protein - the Protein object set at the found conformation and with updated properties according to the performed moves.
depth_first_bnb(protein)
Finds the most optimal conformation using a depth-first branch-and-bound algorithm.
Does not reset the Protein properties beforehand!
Parameters:
* protein - Protein: the Protein object to fold.
Returns:
* Protein - the Protein object set at the found conformation and with updated properties according to the performed moves.
beam_search(protein, beam_width=-1)
Finds a best-effort conformation using a beam search algorithm.
Does not reset the Protein properties beforehand!
Parameters:
* protein - Protein: the Protein object to fold.
* beam_width - int (optional): the beam width to use, where -1 indicates traversal of the entire search space.
Returns:
* Protein - the Protein object set at the found conformation and with updated properties according to the performed moves.

AminoAcid - core#

Protein objects use AminoAcid objects internally to keep track of the chain on the grid. AminoAcid objects have read-only properties, which are consulted internally by the Protein object. The AminoAcid class can be imported directly from prospr without specifying a submodule, e.g.

from prospr import AminoAcid

AminoAcid Properties#

When using the Python package, each property can be directly called as an attribute. If the C++ core is used, the property can be accessed using a method. Each property is described below with the Python and C++ syntax for accessing them.

type
Type of the AminoAcid.

Python

.type

C++

.get_type()

Return type

str
index
Index of the AminoAcid within the Protein’s sequence.

Python

.index

C++

.get_index()

Return type

int
prev_move
Move to perform in order to get to the previous AminoAcid in the chain.

Python

.prev_move

C++

.get_prev_move()

Return type

int
next_move
Move to perform in order to get to the next AminoAcid in the chain.

Python

.next_move

C++

.get_next_move()

Return type

int

Datasets#

Prospr provides many datasets. Each can be imported from the prospr.datasets submodule, e.g.

from prospr.datasets import load_vanEck250
load_vanEck250(length=10)
Loads the vanEck250 dataset containing 250 proteins per length, with lengths from 10 till 100.
Parameters:
* length - int (optional): the length of the protein sequences to load.
Returns:
* DataFrame - a Pandas DataFrame containing the protein sequences.
load_vanEck1000(length=10)
Loads the vanEck1000 dataset containing 1000 proteins per length, with lengths from 10 till 100.
Parameters:
* length - int (optional): the length of the protein sequences to load.
Returns:
* DataFrame - a Pandas DataFrame containing the protein sequences.
load_vanEck_hratio(length=25, hratio=0.1)
Loads the vanEck_hratio dataset containing proteins per length and hratio combination.
Parameters:
* length - int (optional): the length of the protein sequences to load.
* hratio - float (optional): the hratio upperbound of the hratio interval to use.
Returns:
* DataFrame - a Pandas DataFrame containing the protein sequences.

Helpers#

The helpers submodule contains functions used internally. These functions are very specific for their internal use, but maybe of use to you as well. Each can be imported from the prospr.helpers submodule, e.g.

from prospr.helpers import get_ordered_positions
get_ordered_positions(protein)
Returns a Numpy ndarray with the positions of all placed amino acids in order of placement.
Parameters:
* protein - Protein: the Protein object to fetch the positions from.
Returns:
* ndarray - a Numpy ndarray with the positions and types of all placed amino acids and in order of placement.
E.g. [[[0, 0], “H”], [[0, 1], “P”], [[[1, 1], “P”], [[[1, 0], “H”]]
get_scoring_aminos(protein)
Returns a dictionary with the positions of all placed amino acids from the given protein that may score points.
Parameters:
* protein - Protein: the Protein object to compute the scoring amino acids from.
Returns:
* dict - a dictionary mapping the locations of all amino acids that may score points to their previous move and next move.
E.g. {[1, 0]: [-1, 2]}
get_scoring_pairs(protein)
Returns a Numpy ndarray with arrays containing tuples of the positions of amino acid pairs from the given protein that score points.
Parameters:
* protein - Protein: the Protein object to compute the scoring pairs from.
Returns:
* ndarray - a Numpy ndarray with arrays containing tuples of the positions of amino acid pairs from the given protein that score points.
E.g. [[(0, 0), (0, 1)], [(0, 0), (-1, 0)]]

Protein - core#

The Protein class is the cornerstone of Prospr. An instance tracks many properties and has methods to alter the Protein’s state. First, an overview will be made of all the properties and how to reach them. Second, all the non-property bounded methods will be given. The Protein class can be imported directly from prospr without specifying a submodule, e.g.

from prospr import Protein

Protein Properties#

When using the Python package, each property can be directly called as an attribute. If the C++ core is used, the property can be accessed using a method. Each property is described below with the Python and C++ syntax for accessing them.

bond_values
The ways to form bonds and their stability.

Python

.bond_values

C++

.get_bond_values()

Return type

Dict[str, int]
changes
The number of amino acids placed so far.

Python

.changes

C++

.get_changes()

Return type

int
cur_len
The length of the current conformation.

Python

.cur_len

C++

.get_cur_len()

Return type

int
dim
The maximum dimension in which the Protein can fold.

Python

.dim

C++

.get_dim()

Return type

int
h_idxs
The indexes of the “H” amino acids in the sequence.

Python

.h_idxs

C++

.get_h_idxs()

Return type

List[int]
last_move
The last performed move.

Python

.last_move

C++

.get_last_move()

Return type

List[int]
last_pos
The position of the amimo acid at the end of the current conformation.

Python

.last_pos

C++

.get_last_pos()

Return type

List[int]
max_weigths
For each amino acid, the maximum value a bond can make.

Python

.max_weigths

C++

.get_max_weigths()

Return type

List[int]
score
The score of the current conformation.

Python

.score

C++

.get_score()

Return type

int
sequence
The amino acid sequence of the Protein.

Python

.sequence

C++

.get_sequence()

Return type

str

Methods#

The Protein class knows many methods to interact with a protein. The Python package and C++ core use the same signatures, so no distinction is made in the references below.

.get_amino(position)
Returns a list with the amino acid index and next move of the amino acid placed at the given position.
Parameters:
* position - List[int]: position of the amino acid.
Returns:
* List[int] - a list with the amino acid’s index and next move.
E.g. [0, 1]
.get_bonds()
Returns a list of amino acid index pairs that are bonding.
Parameters:
* None
Returns:
* List[Tuple[int,int]] - a list of tuples with two amino acid indexes that bond.
E.g. [(0, 9), (2, 9), (9, 2), (9, 0)]
.hash_fold()
Returns a list of moves representing the current conformation.
Parameters:
* None
Returns:
* List[int] - a list of moves.
E.g. [1, 2, -1]
.is_hydro(index)
Returns if the amino acid at the given index is an H.
Parameters:
* index - int: index of the amino acid.
Returns:
* bool - A boolean indicating if the amino acid is an H.
.is_valid(move)
Returns if the given move is a valid next move.
Parameters:
* move - int: possible next move to perform.
Returns:
* bool - A boolean indicating if the given move is valid.
.place_amino(move, track=True)
Places the next amino acid in the given direction.
Parameters:
* move - int: direction to place the next amino acid.
* track - bool (optional): set to True if the move should be tracked as a change.
Returns:
* None
.remove_amino()
Removes the previously placed amino acid.
Parameters:
* None
Returns:
* None
.reset()
Reset the whole Protein as if it was just created.
Parameters:
* None
Returns:
* None
.reset_conformation()
Reset the placement of amino acids for the given Protein. Also sets the .score property to 0.
Parameters:
* None
Returns:
* None
.set_hash(fold_hash)
Set the conformation to the given sequence of moves.
Parameters:
* fold_hash - List[int]: a list of moves as provided by .hash_fold().
Returns:
* None

Visualize#

Functions from the visualize module of Prospr can be used to illustrate your research. Each function can be imported from the prospr.visualize submodule, e.g.

from prospr.visualize import plot_protein
plot_protein(protein, style=”basic”, ax=None, legend=True, legend_style=”inner”, show=True,
linewidth=2.5, markersize=210, annotate_first=False)
Plots the current set conformation of the given Protein object.
Parameters:
* protein - Protein: a Protein object to plot the conformation of.
* style - str (optional): The figure style to use, either ‘basic’ or ‘paper’.
* ax - Axes (optional): If given, plot the conformation on the given Matplotlib Axes.
* legend - bool (optional): Set to False to disable the legend.
* legend_style - str (optional): The legend style to use, either ‘inner’ or ‘outer’.
* show - bool (optional): Set to False to disable plt.show() call.
* linewidth - float (optional): Line width of the chain.
* markersize - float (optional): Size of the amino acids.
* annotate_first - float (optional): Set to True to highlight first amino acid with a color.
Returns:
* None