def interval_tree(self):
if self._interval_tree is None:
self._interval_tree = IntervalTreeNode.build(self.clusters)
if self._interval_tree is None:
raise ValueError(
"Could not build intervals for peak retrieval with %d clusters" % len(
self.clusters), self)
return self._interval_tree
def test_build(self):
# generate many redundant intervals to force
# the interval tree to branch out.
intervals = self.make_intervals() * 30
ivt = IntervalTreeNode.build(intervals)
assert ivt.start == 3
assert ivt.end == 36
assert ivt.left is not None
assert ivt.left.start == 3
assert ivt.left.end == 12
assert ivt.right is not None
assert ivt.right.end == 36
assert ivt.right.start == 22
def test_build(self):
# generate many redundant intervals to force
# the interval tree to branch out.
intervals = self.make_intervals() * 30
ivt = IntervalTreeNode.build(intervals)
assert ivt.start == 3
assert ivt.end == 36
assert ivt.left is not None
assert ivt.left.start == 3
assert ivt.left.end == 12
assert ivt.right is not None
assert ivt.right.end == 36
assert ivt.right.start == 22
def find_solution_for(self, feature):
feature_node = self.nodes[feature]
tree = self.interval_tree
if tree is None:
tree = IntervalTreeNode.build(self.clusters)
if tree is None:
raise ValueError(
"Could not build intervals for peak retrieval with %d clusters" % len(
self.clusters))
clusters = tree.contains_point(feature.mz)
if len(clusters) == 0:
return self._find_fuzzy_solution_for(feature)
best_fits = [cluster.disjoint_best_fits() for cluster in clusters]
acc = []
for fits in best_fits:
acc.extend(fits)
best_fits = acc
common = tuple(set(best_fits) & set(feature_node.links))
if len(common) > 1 or len(common) == 0:
if len(common) > 1:
warnings.warn("Too many solutions exist for %r" % feature)
# If there were no fits for this peak, then it may be that this peak
# was not included in a fit. Try to find the nearest solution.
i = 0
err = float('inf')
for j, case in enumerate(best_fits):
case_err = abs(case.monoisotopic_feature.mz - feature.mz)
if case_err < err:
i = j
err = case_err
fit = best_fits[i]
else:
fit = common[0]
return self._solution_map[fit]
def split_protein(self, protein_obj, sites=None):
if sites is None:
sites = []
if not sites:
return
seen = set()
sites_seen = set()
peptides = protein_obj.peptides.all()
peptide_intervals = IntervalTreeNode.build(map(PeptideInterval, peptides))
for site in sites:
overlap_region = peptide_intervals.contains_point(site - 1)
spanned_intervals = IntervalTreeNode.build(overlap_region)
# No spanned peptides. May be caused by regions of protein which digest to peptides
# of unacceptable size.
if spanned_intervals is None:
continue
lo = spanned_intervals.start
hi = spanned_intervals.end
# Get the set of all sites spanned by any peptide which spans the current query site
spanned_sites = [s for s in sites if lo <= s <= hi]
for i in range(1, len(spanned_sites) + 1):
for split_sites in itertools.combinations(spanned_sites, i):
site_key = frozenset(split_sites)
if site_key in sites_seen:
continue
sites_seen.add(site_key)
spanning_peptides_query = spanned_intervals.contains_point(split_sites[0])
for site_j in split_sites[1:]:
spanning_peptides_query = [
sp for sp in spanning_peptides_query if site_j in sp
]
spanning_peptides = []
for sp in spanning_peptides_query:
spanning_peptides.extend(sp)
for peptide in spanning_peptides:
adjusted_sites = [0] + [s - peptide.start_position for s in split_sites] + [
peptide.sequence_length]
for j in range(len(adjusted_sites) - 1):
begin, end = adjusted_sites[j], adjusted_sites[j + 1]
if end - begin < self.min_length:
continue
start_position = begin + peptide.start_position
end_position = end + peptide.start_position
if (start_position, end_position) in seen:
continue
else:
seen.add((start_position, end_position))
for modified_peptide, n_variable_modifications in self._permuted_peptides(
peptide.base_peptide_sequence[begin:end]):
inst = Peptide(
base_peptide_sequence=str(peptide.base_peptide_sequence[begin:end]),
modified_peptide_sequence=str(modified_peptide),
count_missed_cleavages=peptide.count_missed_cleavages,
count_variable_modifications=n_variable_modifications,
sequence_length=len(modified_peptide),
start_position=start_position,
end_position=end_position,
calculated_mass=modified_peptide.mass,
formula=formula(modified_peptide.total_composition()),
protein_id=protein_obj.id)
inst.hypothesis_id = protein_obj.hypothesis_id
inst.peptide_score = 0
inst.peptide_score_type = 'null_score'
n_glycosites = n_glycan_sequon_sites(
inst, protein_obj)
o_glycosites = o_glycan_sequon_sites(inst, protein_obj)
gag_glycosites = gag_sequon_sites(inst, protein_obj)
inst.count_glycosylation_sites = len(n_glycosites)
inst.n_glycosylation_sites = sorted(n_glycosites)
inst.o_glycosylation_sites = sorted(o_glycosites)
inst.gagylation_sites = sorted(gag_glycosites)
yield inst
def __init__(self, features):
self.rt_tree = IntervalTreeNode.build(map(RTFeatureNode, features))
def neutral_mass_point_organizer_callback(contained_intervals):
return IntervalTreeNode.build([
Interval(node.neutral_mass, node.neutral_mass, [node])
for node in contained_intervals
])
def mz_point_organizer_callback(contained_intervals):
return IntervalTreeNode.build(
[Interval(node.mz, node.mz, [node]) for node in contained_intervals])
def __init__(self, features):
self.rt_tree = IntervalTreeNode.build(map(RTFeatureNode, features))
def split_protein(self, protein_obj, sites=None):
if sites is None:
sites = []
if not sites:
return
seen = set()
sites_seen = set()
peptides = protein_obj.peptides.all()
peptide_intervals = IntervalTreeNode.build(map(PeptideInterval, peptides))
for site in sites:
overlap_region = peptide_intervals.contains_point(site - 1)
spanned_intervals = IntervalTreeNode.build(overlap_region)
# No spanned peptides. May be caused by regions of protein which digest to peptides
# of unacceptable size.
if spanned_intervals is None:
continue
lo = spanned_intervals.start
hi = spanned_intervals.end
# Get the set of all sites spanned by any peptide which spans the current query site
spanned_sites = [s for s in sites if lo <= s <= hi]
for i in range(1, len(spanned_sites) + 1):
for split_sites in itertools.combinations(spanned_sites, i):
site_key = frozenset(split_sites)
if site_key in sites_seen:
continue
sites_seen.add(site_key)
spanning_peptides_query = spanned_intervals.contains_point(split_sites[0])
for site_j in split_sites[1:]:
spanning_peptides_query = [
sp for sp in spanning_peptides_query if site_j in sp
]
spanning_peptides = []
for sp in spanning_peptides_query:
spanning_peptides.extend(sp)
for peptide in spanning_peptides:
adjusted_sites = [0] + [s - peptide.start_position for s in split_sites] + [
peptide.sequence_length]
for j in range(len(adjusted_sites) - 1):
begin, end = adjusted_sites[j], adjusted_sites[j + 1]
if end - begin < self.min_length:
continue
start_position = begin + peptide.start_position
end_position = end + peptide.start_position
if (start_position, end_position) in seen:
continue
else:
seen.add((start_position, end_position))
for modified_peptide, n_variable_modifications in self._permuted_peptides(
peptide.base_peptide_sequence[begin:end]):
inst = Peptide(
base_peptide_sequence=str(peptide.base_peptide_sequence[begin:end]),
modified_peptide_sequence=str(modified_peptide),
count_missed_cleavages=peptide.count_missed_cleavages,
count_variable_modifications=n_variable_modifications,
sequence_length=len(modified_peptide),
start_position=start_position,
end_position=end_position,
calculated_mass=modified_peptide.mass,
formula=formula(modified_peptide.total_composition()),
protein_id=protein_obj.id)
inst.hypothesis_id = protein_obj.hypothesis_id
inst.peptide_score = 0
inst.peptide_score_type = 'null_score'
n_glycosites = n_glycan_sequon_sites(
inst, protein_obj)
o_glycosites = o_glycan_sequon_sites(inst, protein_obj)
gag_glycosites = gag_sequon_sites(inst, protein_obj)
inst.count_glycosylation_sites = len(n_glycosites)
inst.n_glycosylation_sites = sorted(n_glycosites)
inst.o_glycosylation_sites = sorted(o_glycosites)
inst.gagylation_sites = sorted(gag_glycosites)
yield inst