File size: 6,175 Bytes
7a49887
 
 
7852434
 
 
 
 
fe64e43
 
7a49887
 
 
 
 
fe64e43
7a49887
fe64e43
7a49887
fe64e43
 
7a49887
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
fe64e43
7a49887
 
 
 
fe64e43
 
7a49887
fe64e43
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: other
tags:
- biology
- RNA
- Torsional
- Angles
pipeline_tag: token-classification
base_model:
- zhihan1996/DNA_bert_3
---
# `RNA-TorsionBERT`

## Model Description

`RNA-TorsionBERT` is a 86.9 MB parameter BERT-based language model that predicts RNA torsional and pseudo-torsional angles from the sequence. 

`RNA-TorsionBERT` is a DNABERT model that was pre-trained on ~4200 RNA structures. 

It provides improvement of [MCQ](https://github.com/tzok/mcq4structures) over the previous state-of-the-art models like 
[SPOT-RNA-1D](https://github.com/jaswindersingh2/SPOT-RNA-1D) or inferred angles from existing methods, on the Test Set (composed of RNA-Puzzles and CASP-RNA).


**Key Features**
* Torsional and Pseudo-torsional angles prediction
* Predict sequences up to 512 nucleotides

## Usage

Get started generating text with `RNA-TorsionBERT` by using the following code snippet:

```python
from transformers import AutoModel, AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("sayby/rna_torsionbert", trust_remote_code=True)
model = AutoModel.from_pretrained("sayby/rna_torsionbert", trust_remote_code=True)

sequence = "ACG CGG GGT GTT"
params_tokenizer = {
    "return_tensors": "pt",
    "padding": "max_length",
    "max_length": 512,
    "truncation": True,
}
inputs = tokenizer(sequence, **params_tokenizer)
output = model(inputs)["logits"]
```

- Please note that it was fine-tuned from a DNABERT-3 model and therefore the tokenizer is the same as the one used for DNABERT. Nucleotide `U` should therefore be replaced by `T` in the input sequence.
- The output is the sinus and the cosine for each angle. The angles are in the following order: `alpha`, `beta`,`gamma`,`delta`,`epsilon`,`zeta`,`chi`,`eta`,`theta`,`eta'`,`theta'`,`v0`,`v1`,`v2`,`v3`,`v4`.

To convert the predictions into angles, you can use the following code snippet:

```python
import transformers
from transformers import AutoModel, AutoTokenizer
import numpy as np
import pandas as pd
from typing import Optional, Dict
import os

os.environ["TOKENIZERS_PARALLELISM"] = "false"

transformers.logging.set_verbosity_error()


BACKBONE = [
    "alpha",
    "beta",
    "gamma",
    "delta",
    "epsilon",
    "zeta",
    "chi",
    "eta",
    "theta",
    "eta'",
    "theta'",
    "v0",
    "v1",
    "v2",
    "v3",
    "v4",
]


class RNATorsionBERTHelper:
    def __init__(self):
        self.model_name = "sayby/rna_torsionbert"
        self.tokenizer = AutoTokenizer.from_pretrained(
            self.model_name, trust_remote_code=True
        )
        self.params_tokenizer = {
            "return_tensors": "pt",
            "padding": "max_length",
            "max_length": 512,
            "truncation": True,
        }
        self.model = AutoModel.from_pretrained(self.model_name, trust_remote_code=True)

    def predict(self, sequence: str):
        sequence_tok = self.convert_raw_sequence_to_k_mers(sequence)
        inputs = self.tokenizer(sequence_tok, **self.params_tokenizer)
        outputs = self.model(inputs)["logits"]
        outputs = self.convert_sin_cos_to_angles(
            outputs.cpu().detach().numpy(), inputs["input_ids"]
        )
        output_angles = self.convert_logits_to_dict(
            outputs[0, :], inputs["input_ids"][0, :].cpu().detach().numpy()
        )
        output_angles.index = list(sequence)[:-2]  # Because of the 3-mer representation
        return output_angles

    def convert_raw_sequence_to_k_mers(self, sequence: str, k_mers: int = 3):
        """
        Convert a raw RNA sequence into sequence readable for the tokenizer.
        It converts the sequence into k-mers, and replace U by T
        :return: input readable by the tokenizer
        """
        sequence = sequence.upper().replace("U", "T")
        k_mers_sequence = [
            sequence[i : i + k_mers]
            for i in range(len(sequence))
            if len(sequence[i : i + k_mers]) == k_mers
        ]
        return " ".join(k_mers_sequence)

    def convert_sin_cos_to_angles(
        self, output: np.ndarray, input_ids: Optional[np.ndarray] = None
    ):
        """
        Convert the raw predictions of the RNA-TorsionBERT into angles.
        It converts the cos and sinus into angles using:
            alpha = arctan(sin(alpha)/cos(alpha))
        :param output: Dictionary with the predictions of the RNA-TorsionBERT per angle
        :param input_ids: the input_ids of the RNA-TorsionBERT. It allows to only select the of the sequence,
            and not the special tokens.
        :return: a np.ndarray with the angles for the sequence
        """
        if input_ids is not None:
            output[
                (input_ids == 0)
                | (input_ids == 2)
                | (input_ids == 3)
                | (input_ids == 4)
            ] = np.nan
        pair_indexes, impair_indexes = np.arange(0, output.shape[-1], 2), np.arange(
            1, output.shape[-1], 2
        )
        sin, cos = output[:, :, impair_indexes], output[:, :, pair_indexes]
        tan = np.arctan2(sin, cos)
        angles = np.degrees(tan)
        return angles

    def convert_logits_to_dict(self, output: np.ndarray, input_ids: np.ndarray) -> Dict:
        """
        Convert the raw predictions into dictionary format.
        It removes the special tokens and only keeps the predictions for the sequence.
        :param output: predictions from the models in angles
        :param input_ids: input ids from the tokenizer
        :return: a dictionary with the predictions for each angle
        """
        index_start, index_end = (
            np.where(input_ids == 2)[0][0],
            np.where(input_ids == 3)[0][0],
        )
        output_non_pad = output[index_start + 1 : index_end, :]
        output_angles = {
            angle: output_non_pad[:, angle_index]
            for angle_index, angle in enumerate(BACKBONE)
        }
        out = pd.DataFrame(output_angles)
        return out


if __name__ == "__main__":
    sequence = "AGGGCUUUAGUCUUUGGAG"
    rna_torsionbert_helper = RNATorsionBERTHelper()
    output_angles = rna_torsionbert_helper.predict(sequence)
    print(output_angles)
```