codonalignment.py

# Copyright 2013 by Zheng Ruan (zruan1991@gmail.com).
# All rights reserved.
# This code is part of the Biopython distribution and governed by its
# license.  Please see the LICENSE file that should have been included
# as part of this package.
"""Code for dealing with Codon Alignment.

CodonAlignment class is inherited from MultipleSeqAlignment class. This is
the core class to deal with codon alignment in biopython.
"""

import warnings

from Bio.Align import MultipleSeqAlignment
from Bio.SeqRecord import SeqRecord
from Bio.Data import CodonTable
from Bio import BiopythonWarning


from Bio.codonalign.codonseq import _get_codon_list, CodonSeq, cal_dn_ds
from Bio.codonalign.chisq import chisqprob


class CodonAlignment(MultipleSeqAlignment):
    """Codon Alignment class that inherits from MultipleSeqAlignment.

    >>> from Bio.SeqRecord import SeqRecord
    >>> a = SeqRecord(CodonSeq("AAAACGTCG"), id="Alpha")
    >>> b = SeqRecord(CodonSeq("AAA---TCG"), id="Beta")
    >>> c = SeqRecord(CodonSeq("AAAAGGTGG"), id="Gamma")
    >>> print(CodonAlignment([a, b, c]))
    CodonAlignment with 3 rows and 9 columns (3 codons)
    AAAACGTCG Alpha
    AAA---TCG Beta
    AAAAGGTGG Gamma

    """

    def __init__(self, records="", name=None):
        """Initialize the class."""
        MultipleSeqAlignment.__init__(self, records)

        # check the type of the alignment to be nucleotide
        for rec in self:
            if not isinstance(rec.seq, CodonSeq):
                raise TypeError(
                    "CodonSeq objects are expected in each SeqRecord in CodonAlignment"
                )

        if self.get_alignment_length() % 3 != 0:
            raise ValueError(
                "Alignment length is not a multiple of "
                "three (i.e. a whole number of codons)"
            )

    def __str__(self):
        """Return a multi-line string summary of the alignment.

        This output is indicated to be readable, but large alignment
        is shown truncated. A maximum of 20 rows (sequences) and
        60 columns (20 codons) are shown, with the record identifiers.
        This should fit nicely on a single screen. e.g.

        """
        rows = len(self._records)
        lines = [
            "CodonAlignment with %i rows and %i columns (%i codons)"
            % (rows, self.get_alignment_length(), self.get_aln_length())
        ]

        if rows <= 60:
            lines.extend([self._str_line(rec, length=60) for rec in self._records])
        else:
            lines.extend([self._str_line(rec, length=60) for rec in self._records[:18]])
            lines.append("...")
            lines.append(self._str_line(self._records[-1], length=60))
        return "\n".join(lines)

    def __getitem__(self, index):
        """Return a CodonAlignment object for single indexing."""
        if isinstance(index, int):
            return self._records[index]
        elif isinstance(index, slice):
            return CodonAlignment(self._records[index])
        elif len(index) != 2:
            raise TypeError("Invalid index type.")
        # Handle double indexing
        row_index, col_index = index
        if isinstance(row_index, int):
            return self._records[row_index][col_index]
        elif isinstance(col_index, int):
            return "".join(str(rec[col_index]) for rec in self._records[row_index])
        else:
            return MultipleSeqAlignment(
                rec[col_index] for rec in self._records[row_index]
            )

    def __add__(self, other):
        """Combine two codonalignments with the same number of rows by adding them.

        The method also allows to combine a CodonAlignment object with a
        MultipleSeqAlignment object. The following rules apply:

            * CodonAlignment + CodonAlignment -> CodonAlignment
            * CodonAlignment + MultipleSeqAlignment -> MultipleSeqAlignment
        """
        if isinstance(other, CodonAlignment):
            if len(self) != len(other):
                raise ValueError(
                    "When adding two alignments they must have the same length"
                    " (i.e. same number or rows)"
                )
            warnings.warn(
                "Please make sure the two CodonAlignment objects are sharing the same codon table. This is not checked by Biopython.",
                BiopythonWarning,
            )
            merged = (
                SeqRecord(seq=CodonSeq(left.seq + right.seq))
                for left, right in zip(self, other)
            )
            return CodonAlignment(merged)
        elif isinstance(other, MultipleSeqAlignment):
            if len(self) != len(other):
                raise ValueError(
                    "When adding two alignments they must have the same length"
                    " (i.e. same number or rows)"
                )
            return self.toMultipleSeqAlignment() + other
        else:
            raise TypeError(
                "Only CodonAlignment or MultipleSeqAlignment object can be"
                f" added with a CodonAlignment object. {object(other)} detected."
            )

    def get_aln_length(self):
        """Get alignment length."""
        return self.get_alignment_length() // 3

    def toMultipleSeqAlignment(self):
        """Convert the CodonAlignment to a MultipleSeqAlignment.

        Return a MultipleSeqAlignment containing all the
        SeqRecord in the CodonAlignment using Seq to store
        sequences
        """
        alignments = [SeqRecord(rec.seq.toSeq(), id=rec.id) for rec in self._records]
        return MultipleSeqAlignment(alignments)

    def get_dn_ds_matrix(self, method="NG86", codon_table=None):
        """Available methods include NG86, LWL85, YN00 and ML.

        Argument:
         - method       - Available methods include NG86, LWL85, YN00 and ML.
         - codon_table  - Codon table to use for forward translation.

        """
        from Bio.Phylo.TreeConstruction import DistanceMatrix as DM

        if codon_table is None:
            codon_table = CodonTable.generic_by_id[1]
        names = [i.id for i in self._records]
        size = len(self._records)
        dn_matrix = []
        ds_matrix = []
        for i in range(size):
            dn_matrix.append([])
            ds_matrix.append([])
            for j in range(i + 1):
                if i != j:
                    dn, ds = cal_dn_ds(
                        self._records[i],
                        self._records[j],
                        method=method,
                        codon_table=codon_table,
                    )
                    dn_matrix[i].append(dn)
                    ds_matrix[i].append(ds)
                else:
                    dn_matrix[i].append(0.0)
                    ds_matrix[i].append(0.0)
        dn_dm = DM(names, matrix=dn_matrix)
        ds_dm = DM(names, matrix=ds_matrix)
        return dn_dm, ds_dm

    def get_dn_ds_tree(
        self, dn_ds_method="NG86", tree_method="UPGMA", codon_table=None
    ):
        """Construct dn tree and ds tree.

        Argument:
         - dn_ds_method - Available methods include NG86, LWL85, YN00 and ML.
         - tree_method  - Available methods include UPGMA and NJ.

        """
        from Bio.Phylo.TreeConstruction import DistanceTreeConstructor

        if codon_table is None:
            codon_table = CodonTable.generic_by_id[1]
        dn_dm, ds_dm = self.get_dn_ds_matrix(
            method=dn_ds_method, codon_table=codon_table
        )
        dn_constructor = DistanceTreeConstructor()
        ds_constructor = DistanceTreeConstructor()
        if tree_method == "UPGMA":
            dn_tree = dn_constructor.upgma(dn_dm)
            ds_tree = ds_constructor.upgma(ds_dm)
        elif tree_method == "NJ":
            dn_tree = dn_constructor.nj(dn_dm)
            ds_tree = ds_constructor.nj(ds_dm)
        else:
            raise RuntimeError(
                f"Unknown tree method ({tree_method})."
                " Only NJ and UPGMA are accepted."
            )
        return dn_tree, ds_tree

    @classmethod
    def from_msa(cls, align):
        """Convert a MultipleSeqAlignment to CodonAlignment.

        Function to convert a MultipleSeqAlignment to CodonAlignment.
        It is the user's responsibility to ensure all the requirement
        needed by CodonAlignment is met.
        """
        rec = [SeqRecord(CodonSeq(str(i.seq)), id=i.id) for i in align._records]
        return cls(rec)


def mktest(codon_alns, codon_table=None, alpha=0.05):
    """McDonald-Kreitman test for neutrality.

    Implement the McDonald-Kreitman test for neutrality (PMID: 1904993)
    This method counts changes rather than sites
    (http://mkt.uab.es/mkt/help_mkt.asp).

    Arguments:
     - codon_alns  - list of CodonAlignment to compare (each
       CodonAlignment object corresponds to gene sampled from a species)

    Return the p-value of test result.
    """
    import copy

    if codon_table is None:
        codon_table = CodonTable.generic_by_id[1]
    if not all(isinstance(i, CodonAlignment) for i in codon_alns):
        raise TypeError("mktest accepts CodonAlignment list.")
    codon_aln_len = [i.get_alignment_length() for i in codon_alns]
    if len(set(codon_aln_len)) != 1:
        raise RuntimeError(
            "CodonAlignment object for mktest should be of equal length."
        )
    codon_num = codon_aln_len[0] // 3
    # prepare codon_dict (taking stop codon as an extra amino acid)
    codon_dict = copy.deepcopy(codon_table.forward_table)
    for stop in codon_table.stop_codons:
        codon_dict[stop] = "stop"
    # prepare codon_lst
    codon_lst = []
    for codon_aln in codon_alns:
        codon_lst.append([])
        for i in codon_aln:
            codon_lst[-1].append(_get_codon_list(i.seq))
    codon_set = []
    for i in range(codon_num):
        uniq_codons = []
        for j in codon_lst:
            uniq_codon = {k[i] for k in j}
            uniq_codons.append(uniq_codon)
        codon_set.append(uniq_codons)
    syn_fix, nonsyn_fix, syn_poly, nonsyn_poly = 0, 0, 0, 0
    G, nonsyn_G = _get_codon2codon_matrix(codon_table=codon_table)
    for i in codon_set:
        all_codon = i[0].union(*i[1:])
        if "-" in all_codon or len(all_codon) == 1:
            continue
        fix_or_not = all(len(k) == 1 for k in i)
        if fix_or_not:
            # fixed
            nonsyn_subgraph = _get_subgraph(all_codon, nonsyn_G)
            subgraph = _get_subgraph(all_codon, G)
            this_non = _count_replacement(all_codon, nonsyn_subgraph)
            this_syn = _count_replacement(all_codon, subgraph) - this_non
            nonsyn_fix += this_non
            syn_fix += this_syn
        else:
            # not fixed
            nonsyn_subgraph = _get_subgraph(all_codon, nonsyn_G)
            subgraph = _get_subgraph(all_codon, G)
            this_non = _count_replacement(all_codon, nonsyn_subgraph)
            this_syn = _count_replacement(all_codon, subgraph) - this_non
            nonsyn_poly += this_non
            syn_poly += this_syn
    return _G_test([syn_fix, nonsyn_fix, syn_poly, nonsyn_poly])


def _get_codon2codon_matrix(codon_table):
    """Get codon codon substitution matrix (PRIVATE).

    Elements in the matrix are number of synonymous and nonsynonymous
    substitutions required for the substitution.
    """
    base_tuple = ("A", "T", "C", "G")
    codons = [
        i
        for i in list(codon_table.forward_table.keys()) + codon_table.stop_codons
        if "U" not in i
    ]
    # set up codon_dict considering stop codons
    codon_dict = codon_table.forward_table
    for stop in codon_table.stop_codons:
        codon_dict[stop] = "stop"
    # count site
    num = len(codons)
    G = {}  # graph for substitution
    nonsyn_G = {}  # graph for nonsynonymous substitution
    graph = {}
    graph_nonsyn = {}
    for i, codon in enumerate(codons):
        graph[codon] = {}
        graph_nonsyn[codon] = {}
        for p, b in enumerate(codon):
            for j in base_tuple:
                tmp_codon = codon[0:p] + j + codon[p + 1 :]
                if codon_dict[codon] != codon_dict[tmp_codon]:
                    graph_nonsyn[codon][tmp_codon] = 1
                    graph[codon][tmp_codon] = 1
                else:
                    if codon != tmp_codon:
                        graph_nonsyn[codon][tmp_codon] = 0.1
                        graph[codon][tmp_codon] = 1
    for codon1 in codons:
        nonsyn_G[codon1] = {}
        G[codon1] = {}
        for codon2 in codons:
            if codon1 == codon2:
                nonsyn_G[codon1][codon2] = 0
                G[codon1][codon2] = 0
            else:
                nonsyn_G[codon1][codon2] = _dijkstra(graph_nonsyn, codon1, codon2)
                G[codon1][codon2] = _dijkstra(graph, codon1, codon2)
    return G, nonsyn_G


def _dijkstra(graph, start, end):
    """Dijkstra's algorithm Python implementation (PRIVATE).

    Algorithm adapted from
    http://thomas.pelletier.im/2010/02/dijkstras-algorithm-python-implementation/.
    However, an obvious bug in::

        if D[child_node] >(<) D[node] + child_value:

    is fixed.
    This function will return the distance between start and end.

    Arguments:
     - graph: Dictionary of dictionary (keys are vertices).
     - start: Start vertex.
     - end: End vertex.

    Output:
       List of vertices from the beginning to the end.

    """
    D = {}  # Final distances dict
    P = {}  # Predecessor dict
    # Fill the dicts with default values
    for node in graph.keys():
        D[node] = 100  # Vertices are unreachable
        P[node] = ""  # Vertices have no predecessors
    D[start] = 0  # The start vertex needs no move
    unseen_nodes = list(graph.keys())  # All nodes are unseen
    while len(unseen_nodes) > 0:
        # Select the node with the lowest value in D (final distance)
        shortest = None
        node = ""
        for temp_node in unseen_nodes:
            if shortest is None:
                shortest = D[temp_node]
                node = temp_node
            elif D[temp_node] < shortest:
                shortest = D[temp_node]
                node = temp_node
        # Remove the selected node from unseen_nodes
        unseen_nodes.remove(node)
        # For each child (ie: connected vertex) of the current node
        for child_node, child_value in graph[node].items():
            if D[child_node] > D[node] + child_value:
                D[child_node] = D[node] + child_value
                # To go to child_node, you have to go through node
                P[child_node] = node
        if node == end:
            break
    # Set a clean path
    path = []
    # We begin from the end
    node = end
    distance = 0
    # While we are not arrived at the beginning
    while not (node == start):
        if path.count(node) == 0:
            path.insert(0, node)  # Insert the predecessor of the current node
            node = P[node]  # The current node becomes its predecessor
        else:
            break
    path.insert(0, start)  # Finally, insert the start vertex
    for i in range(len(path) - 1):
        distance += graph[path[i]][path[i + 1]]
    return distance


def _count_replacement(codon_set, G):
    """Count replacement needed for a given codon_set (PRIVATE)."""
    from math import floor

    if len(codon_set) == 1:
        return 0, 0
    elif len(codon_set) == 2:
        codons = list(codon_set)
        return floor(G[codons[0]][codons[1]])
    else:
        codons = list(codon_set)
        return _prim(G)


def _prim(G):
    """Prim's algorithm to find minimum spanning tree (PRIVATE).

    Code is adapted from
    http://programmingpraxis.com/2010/04/09/minimum-spanning-tree-prims-algorithm/
    """
    from math import floor
    from collections import defaultdict
    from heapq import heapify, heappop, heappush

    nodes = []
    edges = []
    for i in G.keys():
        nodes.append(i)
        for j in G[i]:
            if (i, j, G[i][j]) not in edges and (j, i, G[i][j]) not in edges:
                edges.append((i, j, G[i][j]))
    conn = defaultdict(list)
    for n1, n2, c in edges:
        conn[n1].append((c, n1, n2))
        conn[n2].append((c, n2, n1))
    mst = []  # minimum spanning tree
    used = set(nodes[0])
    usable_edges = conn[nodes[0]][:]
    heapify(usable_edges)
    while usable_edges:
        cost, n1, n2 = heappop(usable_edges)
        if n2 not in used:
            used.add(n2)
            mst.append((n1, n2, cost))
            for e in conn[n2]:
                if e[2] not in used:
                    heappush(usable_edges, e)
    length = 0
    for p in mst:
        length += floor(p[2])
    return length


def _get_subgraph(codons, G):
    """Get the subgraph that contains all codons in list (PRIVATE)."""
    subgraph = {}
    for i in codons:
        subgraph[i] = {}
        for j in codons:
            if i != j:
                subgraph[i][j] = G[i][j]
    return subgraph


def _G_test(site_counts):
    """G test for 2x2 contingency table (PRIVATE).

    Arguments:
     - site_counts - [syn_fix, nonsyn_fix, syn_poly, nonsyn_poly]

    >>> print("%0.6f" % _G_test([17, 7, 42, 2]))
    0.004924
    """
    # TODO:
    #   Apply continuity correction for Chi-square test.
    from math import log

    # from scipy.stats import chi2
    G = 0
    tot = sum(site_counts)
    tot_syn = site_counts[0] + site_counts[2]
    tot_non = site_counts[1] + site_counts[3]
    tot_fix = sum(site_counts[:2])
    tot_poly = sum(site_counts[2:])
    exp = [
        tot_fix * tot_syn / tot,
        tot_fix * tot_non / tot,
        tot_poly * tot_syn / tot,
        tot_poly * tot_non / tot,
    ]
    for obs, ex in zip(site_counts, exp):
        G += obs * log(obs / ex)
    G *= 2
    # return 1-chi2.cdf(G, 1) # only 1 dof for 2x2 table
    return chisqprob(G, 1)


if __name__ == "__main__":
    from Bio._utils import run_doctest

    run_doctest()