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
| import numpy as np
import pickle
from Loader import SignedPairsDataset, get_index_dicts
from Trainer import ERGOLightning
from torch.utils.data import DataLoader
from argparse import Namespace
import torch
import pandas as pd
import os
from os import listdir
from os.path import isfile, join
import sys
def read_input_file(datafile):
amino_acids = [letter for letter in 'ARNDCEQGHILKMFPSTWYV']
all_pairs = []
def invalid(seq):
return pd.isna(seq) or any([aa not in amino_acids for aa in seq])
data = pd.read_csv(datafile)
for index in range(len(data)):
sample={}
sample['tcrb'] = data['CDR3'][index]
sample['peptide'] = data['Antigen'][index]
hla_value = data.get('HLA', [None])[index]
if pd.isna(hla_value):
sample['mhc'] = 'UNK'
else:
sample['mhc'] = f"HLA-{hla_value.replace('*', '')}"
sample['tcra'] = 'UNK'
sample['va'] = 'UNK'
sample['ja'] = 'UNK'
sample['vb'] = 'UNK'
sample['jb'] = 'UNK'
sample['t_cell_type'] = 'UNK'
sample['protein'] = 'UNK'
sample['sign'] = 0
if invalid(sample['tcrb']) or invalid(sample['peptide']):
continue
all_pairs.append(sample)
return all_pairs, data
def load_model(hparams, checkpoint_path):
model = ERGOLightning(hparams)
checkpoint = torch.load(checkpoint_path, map_location='cpu')
model.load_state_dict(checkpoint['state_dict'])
model.eval()
return model
def get_model(dataset):
if dataset == 'vdjdb':
version = '1veajht'
if dataset == 'mcpas':
version = '1meajht'
checkpoint_path = os.path.join('Models', 'version_' + version, 'checkpoints')
files = [f for f in listdir(checkpoint_path) if isfile(join(checkpoint_path, f))]
checkpoint_path = os.path.join(checkpoint_path, files[0])
args_path = os.path.join('Models', 'version_' + version, 'meta_tags.csv')
with open(args_path, 'r') as file:
lines = file.readlines()
args = {}
for line in lines[1:]:
key, value = line.strip().split(',')
if key in ['dataset', 'tcr_encoding_model', 'cat_encoding']:
args[key] = value
else:
args[key] = eval(value)
hparams = Namespace(**args)
checkpoint = checkpoint_path
model = load_model(hparams, checkpoint)
train_pickle = 'Samples/' + model.dataset + '_train_samples.pickle'
test_pickle = 'Samples/' + model.dataset + '_test_samples.pickle'
return model, train_pickle,test_pickle
def get_train_dicts(train_pickle):
with open(train_pickle, 'rb') as handle:
train = pickle.load(handle)
train_dicts = get_index_dicts(train)
return train_dicts
def predict(dataset, test_file):
model, train_file,test_pickle = get_model(dataset)
train_dicts = get_train_dicts(train_file)
test_samples, dataframe = read_input_file(test_file)
test_dataset = SignedPairsDataset(test_samples, train_dicts)
batch_size = 1000
loader = DataLoader(test_dataset, batch_size=batch_size, shuffle=False,
collate_fn=lambda b: test_dataset.collate(b, tcr_encoding=model.tcr_encoding_model,
cat_encoding=model.cat_encoding))
outputs = []
def predict(dataset, test_file):
model, train_file, test_pickle = get_model(dataset)
train_dicts = get_train_dicts(train_file)
test_samples, dataframe = read_input_file(test_file)
test_dataset = SignedPairsDataset(test_samples, train_dicts)
batch_size = 1000
loader = DataLoader(test_dataset, batch_size=batch_size, shuffle=False,
collate_fn=lambda b: test_dataset.collate(b, tcr_encoding=model.tcr_encoding_model,
cat_encoding=model.cat_encoding))
outputs = []
for batch_idx, batch in enumerate(loader):
output = model.validation_step(batch, batch_idx)
if output:
outputs.extend(output['y_hat'].tolist())
if len(outputs) != len(dataframe):
print(f"Length mismatch detected!")
print(f"Outputs length: {len(outputs)}")
print(f"Dataframe length: {len(dataframe)}")
print(f"Sample outputs: {outputs[:5]}")
raise ValueError(f"Length mismatch: outputs ({len(outputs)}) vs dataframe ({len(dataframe)})")
dataframe['Score'] = outputs
return dataframe
if __name__ == '__main__':
df = predict('mcpas','test_neg_part1.csv')
print(df)
df.to_csv('results/2.csv', index=False)
pass
|
