def read_alignment(self,
aln_format="psicov",
max_gap_pos=100,
max_gap_seq=100):
self.msa = io.read_msa(self.alignment_file, aln_format)
if max_gap_seq < 100:
self.msa = gaps.remove_gapped_sequences(self.msa, max_gap_seq)
self.max_gap_seq = max_gap_seq
if max_gap_pos < 100:
self.msa, self.gapped_positions = gaps.remove_gapped_positions(
self.msa, max_gap_pos)
self.max_gap_pos = max_gap_pos
self.N = self.msa.shape[0]
self.L = self.msa.shape[1]
self.diversity = np.sqrt(self.N) / self.L
self.neff_entropy = ccmpred.weighting.get_HHsuite_neff(self.msa)
print(
"{0} is of length L={1} and there are {2} sequences in the alignment."
.format(self.protein, self.L, self.N))
print(
"Alignment has diversity [sqrt(N)/L]={0} and Neff(HHsuite-like)={1}."
.format(np.round(self.diversity, decimals=3),
np.round(self.neff_entropy, decimals=3)))
def plot_alignment_statistics(alignment_file, sample_aln_file, aln_format, max_gap_pos, plot_file):
#read alignment
try:
alignment = io.read_msa(alignment_file, aln_format)
except OSError as e:
print("Problems reading alignment file {0}: {1}!".format(alignment_file, e))
sys.exit(0)
try:
sampled_alignment = io.read_msa(sample_aln_file, aln_format)
except OSError as e:
print("Problems reading alignment file {0}: {1}!".format(sample_aln_file, e))
sys.exit(0)
#Remove positions with > MAX_GAP_POS % gaps
if max_gap_pos < 100:
alignment, gapped_positions = gaps.remove_gapped_positions(alignment, max_gap_pos)
non_gapped_positions = [i for i in range(sampled_alignment.shape[1]) if i not in gapped_positions]
sampled_alignment = np.ascontiguousarray(sampled_alignment[:, non_gapped_positions])
# compute sequence weights for observed sequences
weights = ccmpred.weighting.weights_simple(alignment, 0.8)
# compute observed amino acid frequencies
pseudocounts = PseudoCounts(alignment, weights)
pseudocounts.calculate_frequencies(
'uniform_pseudocounts', 1, 1, remove_gaps=False
)
single_freq_observed, pairwise_freq_observed = pseudocounts.freqs
# compute sequence weights for sampled sequences (usually all sampled sequences obtain weight = 1 )
weights_sampled = ccmpred.weighting.weights_simple(sampled_alignment, 0.8)
# compute sampled amino acid frequencies
pseudocounts = PseudoCounts(sampled_alignment, weights_sampled)
pseudocounts.calculate_frequencies(
'uniform_pseudocounts', 1, 1, remove_gaps=False
)
single_freq_sampled, pairwise_freq_sampled = pseudocounts.freqs
# degap the frequencies (ignore gap frequencies)
single_freq_observed = pseudocounts.degap(single_freq_observed, False)
single_freq_sampled = pseudocounts.degap(single_freq_sampled, False)
pairwise_freq_observed = pseudocounts.degap(pairwise_freq_observed, False)
pairwise_freq_sampled = pseudocounts.degap(pairwise_freq_sampled, False)
# plot
plot.plot_empirical_vs_model_statistics(
single_freq_observed, single_freq_sampled,
pairwise_freq_observed, pairwise_freq_sampled,
plot_file)