def get_alignment(cls, seq1: str, seq2: str, local: bool = True):
"""
Generate an alignment between two sequences
Parameters
----------
seq1: str
The first sequence to be aligned
seq1: str
The second sequence to be aligned
local: bool
If false, a global alignment is performed
(based on the Needleman-Wunsch algorithm),
otherwise a local alignment is performed
(based on the Smith–Waterman algorithm).
(Default: True)
Returns
-------
Alignment
"""
import biotite.sequence as seq
import biotite.sequence.align as align
import numpy as np
# create the default matrix
# TODO add more options for the choice of matrix
matrix = align.SubstitutionMatrix.std_protein_matrix()
alignments = align.align_optimal(
seq.ProteinSequence(seq1),
seq.ProteinSequence(seq2),
matrix,
local=local,
)
alignment = alignments[0]
score = alignment.score
seq_identity = align.get_sequence_identity(alignment)
symbols = align.get_symbols(alignment)
codes = align.get_codes(alignment)
return cls(
alignment=alignment,
metadata={
"score": score,
"sequence_identity": seq_identity,
"symbols": symbols,
"codes": codes,
},
)
def mutual_information_zscore(alignment, n_shuffle=100):
codes = align.get_codes(alignment).T
alph = alignment.sequences[0].alphabet
mi = _mutual_information(codes, alph)
np.random.seed(0)
random_mi = [None] * n_shuffle
for i in range(n_shuffle):
shuffled_codes = _shuffle(codes)
random_mi[i] = _mutual_information(shuffled_codes, alph)
random_mi = np.stack(random_mi)
mean = np.mean(random_mi, axis=0)
std = np.std(random_mi, axis=0)
z_score = (mi - mean) / std
return z_score
# If you are interested in more advanced visualization examples, have a
# look at the :doc:`example gallery <../examples/gallery/index>`.
#
# You can also do some simple analysis on these objects, like
# determining the sequence identity or calculating the score.
# For further custom analysis, it can be convenient to have directly the
# aligned symbos codes instead of the trace.
alignment = alignments[0]
print("Score: ", alignment.score)
print("Recalculated score:", align.score(alignment, matrix=matrix))
print("Sequence identity:", align.get_sequence_identity(alignment))
print("Symbols:")
print(align.get_symbols(alignment))
print("symbols codes:")
print(align.get_codes(alignment))
########################################################################
#
# .. currentmodule:: biotite.sequence.io.fasta
#
# You may ask, why should you recalculate the score, when the score has
# already been directly calculated via :func:`align_optimal()`.
# The answer is, that you might load an alignment from an external
# alignment program as FASTA file using :func:`get_alignment()`.
#
# .. currentmodule:: biotite.sequence.align
#
# If you want to perform a multiple sequence alignment, have a look at
# the :func:`align_multiple()` function or the interfaces to external
# MSA software in the :mod:`biotite.application` subpackage.
similarities[i] = 0
else:
sim = matrix[code1, code2]
# Normalize (range 0.0 - 1.0)
min_sim = np.min(matrix[code1])
max_sim = np.max(matrix[code1])
sim = (sim - min_sim) / (max_sim - min_sim)
similarities[i] = sim
# Delete self-similarity
similarities = np.delete(similarities, seq_i)
return np.average(similarities)
matrix = align.SubstitutionMatrix.std_protein_matrix()
# Get the alignment columns as symbols codes (-1 for gaps)
trace_code = align.get_codes(alignment)
similarities = np.zeros(trace_code.shape)
for i in range(similarities.shape[0]):
for j in range(similarities.shape[1]):
similarities[i, j] = get_average_normalized_similarity(
trace_code, matrix.score_matrix(), i, j)
figure = plt.figure(figsize=(8.0, 3.0))
ax = figure.add_subplot(111)
heatmap = ax.pcolor(similarities, cmap="RdYlGn", vmin=0.0, vmax=1.0)
cbar = figure.colorbar(heatmap)
figure.tight_layout()
########################################################################
# As the plot creates a heatmap field for every alignment column,
# the plot looks quite confusing.
#
# Finally, we predict and plot the secondary structure of the *M1* RNA
# with help from *ViennaRNA* and highlight mismatch position between
# *E. coli* and *S. enterica* *M1*.
app = viennarna.RNAfoldApp(m1_sequence)
app.start()
app.join()
base_pairs = app.get_base_pairs()
app = viennarna.RNAplotApp(base_pairs=base_pairs, length=len(m1_sequence))
app.start()
app.join()
plot_coord = app.get_coordinates()
codes = align.get_codes(best_alignment)
m1_no_gap_codes = codes[codes[:, 0] != -1]
identities = m1_no_gap_codes[0] == m1_no_gap_codes[1]
fig = plt.figure(figsize=(8.0, 8.0))
ax = fig.add_subplot(111)
# Plot base connections
ax.plot(*plot_coord.T, color="black", linewidth=1, zorder=1)
# Plot base pairings
ax.add_collection(
LineCollection([(plot_coord[i], plot_coord[j]) for i, j in base_pairs],
color="silver",
linewidth=1,
zorder=1))
# Plot base markers
ax.scatter(
