File size: 5,798 Bytes
e76b79a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
# Copyright (c) Meta Platforms, Inc. and affiliates.

# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.

from abc import ABC, abstractmethod
from copy import deepcopy
from typing import List, Union

import numpy as np
from biotite.structure import AtomArray

ALL_RESIDUE_TYPES = [
    "A",
    "R",
    "N",
    "D",
    "C",
    "Q",
    "E",
    "G",
    "H",
    "I",
    "L",
    "K",
    "M",
    "F",
    "P",
    "S",
    "T",
    "W",
    "Y",
    "V",
]

RESIDUE_TYPES_WITHOUT_CYSTEINE = deepcopy(ALL_RESIDUE_TYPES)
RESIDUE_TYPES_WITHOUT_CYSTEINE.remove("C")

RESIDUE_TYPES_1to3 = {
    "A": "ALA",
    "R": "ARG",
    "N": "ASN",
    "D": "ASP",
    "C": "CYS",
    "Q": "GLN",
    "E": "GLU",
    "G": "GLY",
    "H": "HIS",
    "I": "ILE",
    "L": "LEU",
    "K": "LYS",
    "M": "MET",
    "F": "PHE",
    "P": "PRO",
    "S": "SER",
    "T": "THR",
    "W": "TRP",
    "Y": "TYR",
    "V": "VAL",
}
RESIDUE_TYPES_3to1 = {v: k for k, v in RESIDUE_TYPES_1to3.items()}


class SequenceSegmentFactory(ABC):
    def __init__(self) -> None:
        pass

    @abstractmethod
    def get(self) -> str:
        pass

    @abstractmethod
    def mutate(self) -> None:
        pass

    @abstractmethod
    def num_mutation_candidates(self) -> int:
        pass


class ConstantSequenceSegment(SequenceSegmentFactory):
    def __init__(self, sequence: str) -> None:
        super().__init__()
        self.sequence = sequence

    def get(self) -> str:
        return self.sequence

    def mutate(self) -> None:
        pass

    def num_mutation_candidates(self) -> int:
        return 0


class FixedLengthSequenceSegment(SequenceSegmentFactory):
    def __init__(
        self, initial_sequence: Union[str, int], disallow_mutations_to_cysteine=True,
    ) -> None:
        super().__init__()
        self.mutation_residue_types = (
            RESIDUE_TYPES_WITHOUT_CYSTEINE
            if disallow_mutations_to_cysteine
            else ALL_RESIDUE_TYPES
        )

        self.sequence = (
            initial_sequence
            if type(initial_sequence) == str
            else random_sequence(
                length=initial_sequence, corpus=self.mutation_residue_types
            )
        )

    def get(self) -> str:
        return self.sequence

    def mutate(self) -> None:
        self.sequence = substitute_one_amino_acid(
            self.sequence, self.mutation_residue_types
        )

    def num_mutation_candidates(self) -> int:
        return len(self.sequence)


def substitute_one_amino_acid(sequence: str, corpus: List[str]) -> str:
    sequence = list(sequence)
    index = np.random.choice(len(sequence))
    sequence[index] = np.random.choice(corpus)
    return "".join(sequence)


def random_sequence(length: int, corpus: List[str]) -> str:
    "Generate a random sequence using amino acids in corpus."

    return "".join([np.random.choice(corpus) for _ in range(length)])


def sequence_from_atomarray(atoms: AtomArray) -> str:
    return "".join(
        [RESIDUE_TYPES_3to1[aa] for aa in atoms[atoms.atom_name == "CA"].res_name]
    )


class VariableLengthSequenceSegment(SequenceSegmentFactory):
    def __init__(
        self,
        initial_sequence: Union[str, int],
        disallow_mutations_to_cysteine=True,
        mutation_operation_probabilities: List[float] = [
            3., # Substitution weight.
            1., # Deletion weight.
            1., # Insertion weight.
        ],
    ) -> None:
        super().__init__()
        self.mutation_residue_types = (
            RESIDUE_TYPES_WITHOUT_CYSTEINE
            if disallow_mutations_to_cysteine
            else ALL_RESIDUE_TYPES
        )

        self.sequence = (
            initial_sequence
            if type(initial_sequence) == str
            else random_sequence(
                length=initial_sequence, corpus=self.mutation_residue_types
            )
        )

        self.mutation_operation_probabilities = np.array(mutation_operation_probabilities)
        self.mutation_operation_probabilities /= self.mutation_operation_probabilities.sum()

    def get(self) -> str:
        return self.sequence

    def mutate(self) -> None:
        mutation_operation = np.random.choice(
            [
                self._mutate_substitution,
                self._mutate_deletion,
                self._mutate_insertion,
            ],
            p=self.mutation_operation_probabilities,
        )
        mutation_operation()

    def _mutate_substitution(self) -> str:
        self.sequence = substitute_one_amino_acid(
            self.sequence, self.mutation_residue_types
        )

    def _mutate_deletion(self) -> str:
        self.sequence = delete_one_amino_acid(self.sequence)

    def _mutate_insertion(self) -> str:
        self.sequence = insert_one_amino_acid(
            self.sequence, self.mutation_residue_types
        )

    def num_mutation_candidates(self) -> int:
        # NOTE(brianhie): This should be `3*len(self.sequence) + 1`,
        # since there are `len(self.sequence)` substitutions and
        # deletions, and `len(self.sequence) + 1` insertions.
        # However, as this is used to weight sequence segments for
        # mutations when combined into a multi-segment program, we
        # just weight by `len(self.sequence)` for now.
        return len(self.sequence)


def delete_one_amino_acid(sequence: str) -> str:
    index = np.random.choice(len(sequence))
    return sequence[:index] + sequence[index + 1 :]


def insert_one_amino_acid(sequence: str, corpus: List[str]) -> str:
    n = len(sequence)
    index = np.random.randint(0, n) if n > 0 else 0
    insertion = np.random.choice(corpus)
    return sequence[:index] + insertion + sequence[index:]