Module asymmetree.seqevolve.Alignment
Constructing the true alignment of simulated sequences.
Expand source code
# -*- coding: utf-8 -*-
"""Constructing the true alignment of simulated sequences."""
import numpy as np
import networkx as nx
__author__ = 'David Schaller'
class AlignmentBuilder:
"""Construction of the true alignment of simulated sequences."""
def __init__(self, tree, sequence_dict, alphabet, include_inner=True):
"""
Parameters
----------
tree : Tree
The tree along which the sequences where simulated.
sequence_dict : dict
A dict containing the TreeNode instances in the tree as keys and
the simulated sequences as values (instances of type EvoSeq).
alphabet
The alphabet of the substitution model that was used.
include_inner : bool, optional
If True, the alignment also contains the sequences of the inner
nodes of the tree. The default is True.
"""
self.tree = tree
self.sequence_dict = sequence_dict
self.include_inner = include_inner
if alphabet[-1] == '-':
self.alphabet = alphabet
else:
self.alphabet = alphabet + '-'
def build(self):
"""Build the true alignment.
Returns
-------
dict
A dict containing the TreeNode instances in the tree as keys and
the str sequences as values that include the necessary gaps.
"""
self._get_preorder()
self._sort_sites()
self._alignment_matrix()
return self._sequences()
def _get_preorder(self):
if self.include_inner:
self.nodes = [v for v in self.tree.preorder()]
else:
self.nodes = [v for v in self.tree.preorder() if not v.children]
self.node_index = {item: index for index, item in enumerate(self.nodes)}
def _sort_sites(self):
G = nx.DiGraph()
for node in self.nodes:
seq = self.sequence_dict[node]
# handle case |seq| = 1 correctly
if len(seq) == 1:
G.add_node(seq[0].site_id)
# add nodes and edges to auxiliary graph
for first, second in seq.element_pairs():
G.add_edge(first.site_id, second.site_id)
self.sites = list(nx.topological_sort(G))
def _alignment_matrix(self):
self.alignment = np.zeros((len(self.nodes), len(self.sites)),
dtype=np.int8)
positions = [self.sequence_dict[v]._first for v in self.nodes]
for j in range(len(self.sites)):
for i in range(len(positions)):
if positions[i] and self.sites[j] == positions[i].site_id:
self.alignment[i, j] = positions[i]._value
positions[i] = positions[i]._next
else:
self.alignment[i, j] = -1
def _sequences(self):
result = {}
for i in range(len(self.nodes)):
sequence = "".join(self.alphabet[x] for x in self.alignment[i, :])
result[self.nodes[i]] = sequence
return resultClasses
class AlignmentBuilder (tree, sequence_dict, alphabet, include_inner=True)-
Construction of the true alignment of simulated sequences.
Parameters
tree:Tree- The tree along which the sequences where simulated.
sequence_dict:dict- A dict containing the TreeNode instances in the tree as keys and the simulated sequences as values (instances of type EvoSeq).
alphabet- The alphabet of the substitution model that was used.
include_inner:bool, optional- If True, the alignment also contains the sequences of the inner nodes of the tree. The default is True.
Expand source code
class AlignmentBuilder: """Construction of the true alignment of simulated sequences.""" def __init__(self, tree, sequence_dict, alphabet, include_inner=True): """ Parameters ---------- tree : Tree The tree along which the sequences where simulated. sequence_dict : dict A dict containing the TreeNode instances in the tree as keys and the simulated sequences as values (instances of type EvoSeq). alphabet The alphabet of the substitution model that was used. include_inner : bool, optional If True, the alignment also contains the sequences of the inner nodes of the tree. The default is True. """ self.tree = tree self.sequence_dict = sequence_dict self.include_inner = include_inner if alphabet[-1] == '-': self.alphabet = alphabet else: self.alphabet = alphabet + '-' def build(self): """Build the true alignment. Returns ------- dict A dict containing the TreeNode instances in the tree as keys and the str sequences as values that include the necessary gaps. """ self._get_preorder() self._sort_sites() self._alignment_matrix() return self._sequences() def _get_preorder(self): if self.include_inner: self.nodes = [v for v in self.tree.preorder()] else: self.nodes = [v for v in self.tree.preorder() if not v.children] self.node_index = {item: index for index, item in enumerate(self.nodes)} def _sort_sites(self): G = nx.DiGraph() for node in self.nodes: seq = self.sequence_dict[node] # handle case |seq| = 1 correctly if len(seq) == 1: G.add_node(seq[0].site_id) # add nodes and edges to auxiliary graph for first, second in seq.element_pairs(): G.add_edge(first.site_id, second.site_id) self.sites = list(nx.topological_sort(G)) def _alignment_matrix(self): self.alignment = np.zeros((len(self.nodes), len(self.sites)), dtype=np.int8) positions = [self.sequence_dict[v]._first for v in self.nodes] for j in range(len(self.sites)): for i in range(len(positions)): if positions[i] and self.sites[j] == positions[i].site_id: self.alignment[i, j] = positions[i]._value positions[i] = positions[i]._next else: self.alignment[i, j] = -1 def _sequences(self): result = {} for i in range(len(self.nodes)): sequence = "".join(self.alphabet[x] for x in self.alignment[i, :]) result[self.nodes[i]] = sequence return resultMethods
def build(self)-
Build the true alignment.
Returns
dict- A dict containing the TreeNode instances in the tree as keys and the str sequences as values that include the necessary gaps.
Expand source code
def build(self): """Build the true alignment. Returns ------- dict A dict containing the TreeNode instances in the tree as keys and the str sequences as values that include the necessary gaps. """ self._get_preorder() self._sort_sites() self._alignment_matrix() return self._sequences()