/
tcrdist_metric.py
153 lines (117 loc) · 3.83 KB
/
tcrdist_metric.py
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
__all__ = [
"AlphaCdr3Tcrdist",
"BetaCdr3Tcrdist",
"Cdr3Tcrdist",
"AlphaTcrdist",
"BetaTcrdist",
"Tcrdist"
]
from abc import abstractmethod
from enum import Enum
from numpy import ndarray
from pandas import DataFrame
from pyrepseq.metric.tcr_metric import TcrMetric
from pyrepseq.metric.tcr_metric.tcrdist.simplified_tcrdist_interface import TcrdistInterface
from scipy.spatial import distance
from typing import Iterable
tcrdist_interface = TcrdistInterface()
class TcrChain(Enum):
ALPHA = 1
BETA = 2
class TcrdistType(Enum):
CDR3 = 1
FULL = 2
class AbstractTcrdist(TcrMetric):
@property
@abstractmethod
def _chains_to_compare(self) -> Iterable[TcrChain]:
pass
@property
@abstractmethod
def _tcrdist_type(self) -> TcrdistType:
pass
def calc_cdist_matrix(
self, anchors: DataFrame, comparisons: DataFrame
) -> ndarray:
super().calc_cdist_matrix(anchors, comparisons)
cdist_matrices = []
if TcrChain.ALPHA in self._chains_to_compare:
alpha_result = self._calc_alpha_cdist(anchors, comparisons)
cdist_matrices.append(alpha_result)
if TcrChain.BETA in self._chains_to_compare:
beta_result = self._calc_beta_cdist(anchors, comparisons)
cdist_matrices.append(beta_result)
return sum(cdist_matrices)
def _calc_alpha_cdist(
self, anchors: DataFrame, comparisons: DataFrame
) -> ndarray:
result = tcrdist_interface.calc_alpha_cdist_matrices(
anchors, comparisons
)
if self._tcrdist_type is TcrdistType.CDR3:
return result["cdr3_a_aa"]
else:
return result["tcrdist"]
def _calc_beta_cdist(
self, anchors: DataFrame, comparisons: DataFrame
) -> ndarray:
result = tcrdist_interface.calc_beta_cdist_matrices(
anchors, comparisons
)
if self._tcrdist_type is TcrdistType.CDR3:
return result["cdr3_b_aa"]
else:
return result["tcrdist"]
def calc_pdist_vector(self, instances: DataFrame) -> ndarray:
super().calc_pdist_vector(instances)
pdist_matrix = self.calc_cdist_matrix(instances, instances)
pdist_vector = distance.squareform(pdist_matrix, checks=False)
return pdist_vector
class AlphaCdr3Tcrdist(AbstractTcrdist):
"""
TcrDist applied to the alpha chain CDR3 sequences.
"""
name = "Alpha CDR3 tcrdist"
distance_bins = range(0, 80 + 1, 2)
_chains_to_compare = [TcrChain.ALPHA]
_tcrdist_type = TcrdistType.CDR3
class BetaCdr3Tcrdist(AbstractTcrdist):
"""
TcrDist applied to the beta chain CDR3 sequences.
"""
name = "Beta CDR3 tcrdist"
distance_bins = range(0, 80 + 1, 2)
_chains_to_compare = [TcrChain.BETA]
_tcrdist_type = TcrdistType.CDR3
class Cdr3Tcrdist(AbstractTcrdist):
"""
TcrDist applied to the alpha and beta chain CDR3 sequences.
"""
name = "CDR3 tcrdist"
distance_bins = range(0, 160 + 1, 2)
_chains_to_compare = [TcrChain.ALPHA, TcrChain.BETA]
_tcrdist_type = TcrdistType.CDR3
class AlphaTcrdist(AbstractTcrdist):
"""
TcrDist applied to the alpha chain.
"""
name = "Alpha tcrdist"
distance_bins = range(0, 300 + 1, 5)
_chains_to_compare = [TcrChain.ALPHA]
_tcrdist_type = TcrdistType.FULL
class BetaTcrdist(AbstractTcrdist):
"""
TcrDist applied to the beta chain.
"""
name = "Beta tcrdist"
distance_bins = range(0, 300 + 1, 5)
_chains_to_compare = [TcrChain.BETA]
_tcrdist_type = TcrdistType.FULL
class Tcrdist(AbstractTcrdist):
"""
TcrDist applied to the alpha and beta chain.
"""
name = "tcrdist"
distance_bins = range(0, 600 + 1, 5)
_chains_to_compare = [TcrChain.ALPHA, TcrChain.BETA]
_tcrdist_type = TcrdistType.FULL