def calc_nonepitope_star_distance(tree, attr='ne_star', seasons = []): ''' calculates the distance at nonepitope sites of any tree node to ref tree -- dendropy tree attr -- the attribute name used to save the result ''' if not hasattr(tree, "nonepitope_star_distance_assigned") or tree.nonepitope_star_distance_assigned==False: setup_epitope_mask() for node in tree.postorder_node_iter(): if len(node.season_tips) and node!=tree.seed_node: if not hasattr(node, 'aa'): node.aa = translate(node.seq) tmp_node = node.parent_node cur_season = min(node.season_tips.keys()) prev_season = seasons[max(0,seasons.index(cur_season)-1)] while True: if tmp_node!=tree.seed_node: if prev_season in tmp_node.season_tips and len(tmp_node.season_tips[prev_season])>0: break else: tmp_node=tmp_node.parent_node else: break if not hasattr(tmp_node, 'aa'): tmp_node.aa = translate(tmp_node.seq) node.__setattr__(attr, nonepitope_distance(node.aa, tmp_node.aa)) else: node.__setattr__(attr, np.nan) tree.nonepitope_star_distance_assigned=True
def calc_rbs_distance(self, tree, attr='rb', ref=None):
'''
calculates the distance at receptor binding sites of any tree node to ref
tree -- dendropy tree
attr -- the attribute name used to save the result
'''
if ref == None:
ref = translate(tree.seed_node.seq)
for node in tree.postorder_node_iter():
if not hasattr(node, 'aa'):
node.aa = translate(node.seq)
node.__setattr__(attr, self.rbs_distance(node.aa, ref))
def calc_nonepitope_distance(self, tree, attr='ne', ref=None):
'''
calculates -1 * distance at nonepitope sites of each node in tree to ref
tree -- dendropy tree
attr -- the attribute name used to save the result
'''
if ref == None:
ref = translate(tree.seed_node.seq)
for node in tree.postorder_node_iter():
if not hasattr(node, 'aa'):
node.aa = translate(node.seq)
distance = self.nonepitope_distance(node.aa, ref)
node.__setattr__(attr, -1 * distance)
def calc_rbs_distance(tree, attr='rb', ref = None): ''' calculates the distance at receptor binding sites of any tree node to ref tree -- dendropy tree attr -- the attribute name used to save the result ''' if not hasattr(tree, "rbs_distance_assigned") or tree.rbs_distance_assigned==False: if ref == None: ref = translate(tree.seed_node.seq) for node in tree.postorder_node_iter(): if not hasattr(node, 'aa'): node.aa = translate(node.seq) node.__setattr__(attr, rbs_distance(node.aa, ref)) tree.rbs_distance_assigned=True
def calc_nonepitope_distance(tree, attr="ne", ref=None):
"""
calculates the distance at nonepitope sites of any tree node to ref
tree -- dendropy tree
attr -- the attribute name used to save the result
"""
if not hasattr(tree, "nonepitope_distance_assigned") or tree.nonepitope_distance_assigned == False:
if ref == None:
ref = translate(tree.seed_node.seq)
for node in tree.postorder_node_iter():
if not hasattr(node, "aa"):
node.aa = translate(node.seq)
node.__setattr__(attr, nonepitope_distance(node.aa, ref))
tree.nonepitope_distance_assigned = True
def calc_epitope_distance(tree, attr='ep', ref = None): ''' calculates the distance at epitope sites of any tree node to ref tree -- dendropy tree attr -- the attribute name used to save the result ''' if not hasattr(tree, "epitope_distance_assigned") or tree.epitope_distance_assigned==False: setup_epitope_mask() if ref == None: ref = translate(tree.seed_node.seq) for node in tree.postorder_node_iter(): if not hasattr(node, 'aa'): node.aa = translate(node.seq) node.__setattr__(attr, epitope_distance(node.aa, ref)) tree.epitope_distance_assigned=True
def calc_tolerance(tree, attr='tol'):
'''
calculates the distance at epitope sites of any tree node to ref
tree -- dendropy tree
attr -- the attribute name used to save the result
'''
from Bio import AlignIO
aa, sites, wt_aa, aa_prob = load_mutational_tolerance()
aln = AlignIO.read('/Users/yujiazhou/Documents/nextflu/H9_nextflu-master/augur/source-data/H1_H3.fasta', 'fasta')
# returns true whenever either of the sequences have a gap
aligned = (np.array(aln)!='-').min(axis=0)
# map alignment positions to sequence positions, subset to aligned amino acids
indices = {}
for seq in aln:
indices[seq.name] = (np.cumsum(np.fromstring(str(seq.seq), dtype='S1')!='-')-1)[aligned]
# make a reduced set of amino-acid probabilities that only contains aligned positions
aa_prob=aa_prob[indices['H1'],:]
# attach another column for non-canonical amino acids
aa_prob = np.hstack((aa_prob, 1e-5*np.ones((aa_prob.shape[0],1))))
if not hasattr(tree, "tolerance_assigned") or tree.tolerance_assigned==False:
for node in tree.postorder_node_iter():
if not hasattr(node, 'aa'):
node.aa = translate(node.seq)
node.__setattr__(attr, calc_fitness_tolerance(node.aa, aa_prob, aa, indices['H3']))
tree.tolerance_assigned=True
def calc_epitope_cross_immunity(self,
tree,
timepoint,
window=2.0,
attr='ep_x'):
'''
calculates the distance at epitope sites to contemporaneous viruses
this should capture cross-immunity of circulating viruses
meant to be used in conjunction with epitope_distance that focuses
on escape from previous human immunity
tree -- dendropy tree
attr -- the attribute name used to save the result
'''
comparison_nodes = []
for node in tree.postorder_node_iter():
if node.is_leaf():
if node.num_date < timepoint and node.num_date > timepoint - window:
comparison_nodes.append(node)
if not hasattr(node, 'np_ep'):
if not hasattr(node, 'aa'):
node.aa = translate(node.seq)
node.np_ep = np.array(list(self.epitope_sites(node.aa)))
print "calculating cross-immunity to " + str(
len(comparison_nodes)) + " comparison nodes"
for node in tree.postorder_node_iter():
mean_distance = 0
count = 0
for comp_node in comparison_nodes:
mean_distance += self.fast_epitope_distance(
node.np_ep, comp_node.np_ep)
count += 1
if count > 0:
mean_distance /= float(count)
node.__setattr__(attr, mean_distance)
def calc_epitope_distance(self, tree, attr='ep', ref=None):
'''
calculates the distance at epitope sites of any tree node to ref
tree -- dendropy tree
attr -- the attribute name used to save the result
'''
for node in tree.postorder_node_iter():
if not hasattr(node, 'np_ep'):
if not hasattr(node, 'aa'):
node.aa = translate(node.seq)
node.np_ep = np.array(list(self.epitope_sites(node.aa)))
if ref == None:
ref = tree.seed_node
for node in tree.postorder_node_iter():
node.__setattr__(attr,
self.fast_epitope_distance(node.np_ep, ref.np_ep))