[!TIP] The MultiMolecule team has confirmed that the provided model and checkpoints are producing the same intermediate representations as the original implementation.

The team releasing ERNIE-RNA did not write this model card for this model so this model card has been written by the MultiMolecule team.

Model Details

ERNIE-RNA is a bert-style model pre-trained on a large corpus of non-coding RNA sequences in a self-supervised fashion. This means that the model was trained on the raw nucleotides of RNA sequences only, with an automatic process to generate inputs and labels from those texts. Please refer to the Training Details section for more information on the training process.

Variants

multimolecule/ernierna: The ERNIE-RNA model pre-trained on non-coding RNA sequences.
multimolecule/ernierna-ss: The ERNIE-RNA model fine-tuned on RNA secondary structure prediction.

Model Specification

Num Layers	Hidden Size	Num Heads	Intermediate Size	Num Parameters (M)	FLOPs (G)	MACs (G)	Max Num Tokens
12	768	12	3072	85.67	96.92	48.34	1024

Usage

The model file depends on the multimolecule library. You can install it using pip:

pip install multimolecule

Direct Use

RNA Secondary Structure Prediction

You can use this model directly with a pipeline for secondary structure prediction:

python

import multimolecule  # you must import multimolecule to register models
from transformers import pipeline

predictor = pipeline("rna-secondary-structure", model="multimolecule/ernierna-ss")
output = predictor("GGUCUCUCUGGUUAGACCAGAUCUGAGCCU")

Downstream Use

Extract Features

Here is how to use this model to get the features of a given sequence in PyTorch:

python

from multimolecule import RnaTokenizer, ErnieRnaModel


tokenizer = RnaTokenizer.from_pretrained("multimolecule/ernierna-ss")
model = ErnieRnaModel.from_pretrained("multimolecule/ernierna-ss")

text = "UAGCUUAUCAGACUGAUGUUG"
input = tokenizer(text, return_tensors="pt")

output = model(**input)

Sequence Classification / Regression

[!NOTE] This model is not fine-tuned for any specific task. You will need to fine-tune the model on a downstream task to use it for sequence classification or regression.

Here is how to use this model as backbone to fine-tune for a sequence-level task in PyTorch:

python

import torch
from multimolecule import RnaTokenizer, ErnieRnaForSequencePrediction


tokenizer = RnaTokenizer.from_pretrained("multimolecule/ernierna-ss")
model = ErnieRnaForSequencePrediction.from_pretrained("multimolecule/ernierna-ss")

text = "UAGCUUAUCAGACUGAUGUUG"
input = tokenizer(text, return_tensors="pt")
label = torch.tensor([1])

output = model(**input, labels=label)

Token Classification / Regression

[!NOTE] This model is not fine-tuned for any specific task. You will need to fine-tune the model on a downstream task to use it for token classification or regression.

Here is how to use this model as backbone to fine-tune for a nucleotide-level task in PyTorch:

python

import torch
from multimolecule import RnaTokenizer, ErnieRnaForTokenPrediction


tokenizer = RnaTokenizer.from_pretrained("multimolecule/ernierna-ss")
model = ErnieRnaForTokenPrediction.from_pretrained("multimolecule/ernierna-ss")

text = "UAGCUUAUCAGACUGAUGUUG"
input = tokenizer(text, return_tensors="pt")
label = torch.randint(2, (len(text), ))

output = model(**input, labels=label)

Contact Classification / Regression

[!NOTE] This model is not fine-tuned for any specific task. You will need to fine-tune the model on a downstream task to use it for contact classification or regression.

Here is how to use this model as backbone to fine-tune for a contact-level task in PyTorch:

python

import torch
from multimolecule import RnaTokenizer, ErnieRnaForContactPrediction


tokenizer = RnaTokenizer.from_pretrained("multimolecule/ernierna-ss")
model = ErnieRnaForContactPrediction.from_pretrained("multimolecule/ernierna-ss")

text = "UAGCUUAUCAGACUGAUGUUG"
input = tokenizer(text, return_tensors="pt")
label = torch.randint(2, (len(text), len(text)))

output = model(**input, labels=label)

Training Details

ERNIE-RNA used Masked Language Modeling (MLM) as the pre-training objective: taking a sequence, the model randomly masks 15% of the tokens in the input then runs the entire masked sentence through the model and has to predict the masked tokens. This is comparable to the Cloze task in language modeling.

Training Data

The ERNIE-RNA model was pre-trained on RNAcentral. RNAcentral is a free, public resource that offers integrated access to a comprehensive and up-to-date set of non-coding RNA sequences provided by a collaborating group of Expert Databases representing a broad range of organisms and RNA types.

ERNIE-RNA applied CD-HIT (CD-HIT-EST) with a cut-off at 100% sequence identity to remove redundancy from the RNAcentral, resulting 25 million unique sequences. Sequences longer than 1024 nucleotides were subsequently excluded. The final dataset contains 20.4 million non-redundant RNA sequences. ERNIE-RNA preprocessed all tokens by replacing "T"s with "S"s.

Note that [RnaTokenizer][multimolecule.RnaTokenizer] will convert "T"s to "U"s for you, you may disable this behaviour by passing replace_T_with_U=False.

Training Procedure

Preprocessing

ERNIE-RNA used masked language modeling (MLM) as the pre-training objective. The masking procedure is similar to the one used in BERT:

Mask rate: 15%
Replacement: <mask> for 80% of masked tokens
Replacement: random token for 10% of masked tokens
Replacement: unchanged token for 10% of masked tokens

Pre-training

The model was trained on 24 NVIDIA V100 GPUs with 32GiB memories.

Learning rate: 1e-4
Learning rate warm-up: 20,000 steps
Weight decay: 0.01

Citation

bibtex

@article {Yin2024.03.17.585376,
	author = {Yin, Weijie and Zhang, Zhaoyu and He, Liang and Jiang, Rui and Zhang, Shuo and Liu, Gan and Zhang, Xuegong and Qin, Tao and Xie, Zhen},
	title = {ERNIE-RNA: An RNA Language Model with Structure-enhanced Representations},
	elocation-id = {2024.03.17.585376},
	year = {2024},
	doi = {10.1101/2024.03.17.585376},
	publisher = {Cold Spring Harbor Laboratory},
	abstract = {With large amounts of unlabeled RNA sequences data produced by high-throughput sequencing technologies, pre-trained RNA language models have been developed to estimate semantic space of RNA molecules, which facilities the understanding of grammar of RNA language. However, existing RNA language models overlook the impact of structure when modeling the RNA semantic space, resulting in incomplete feature extraction and suboptimal performance across various downstream tasks. In this study, we developed a RNA pre-trained language model named ERNIE-RNA (Enhanced Representations with base-pairing restriction for RNA modeling) based on a modified BERT (Bidirectional Encoder Representations from Transformers) by incorporating base-pairing restriction with no MSA (Multiple Sequence Alignment) information. We found that the attention maps from ERNIE-RNA with no fine-tuning are able to capture RNA structure in the zero-shot experiment more precisely than conventional methods such as fine-tuned RNAfold and RNAstructure, suggesting that the ERNIE-RNA can provide comprehensive RNA structural representations. Furthermore, ERNIE-RNA achieved SOTA (state-of-the-art) performance after fine-tuning for various downstream tasks, including RNA structural and functional predictions. In summary, our ERNIE-RNA model provides general features which can be widely and effectively applied in various subsequent research tasks. Our results indicate that introducing key knowledge-based prior information in the BERT framework may be a useful strategy to enhance the performance of other language models.Competing Interest StatementOne patent based on the study was submitted by Z.X. and W.Y., which is entitled as "A Pre-training Approach for RNA Sequences and Its Applications"(application number, no 202410262527.5). The remaining authors declare no competing interests.},
	URL = {https://www.biorxiv.org/content/early/2024/03/17/2024.03.17.585376},
	eprint = {https://www.biorxiv.org/content/early/2024/03/17/2024.03.17.585376.full.pdf},
	journal = {bioRxiv}
}

[!NOTE] The artifacts distributed in this repository are part of the MultiMolecule project. If MultiMolecule supports your research, please cite the MultiMolecule project as follows:

bibtex

@software{chen_2024_12638419,
  author    = {Chen, Zhiyuan and Zhu, Sophia Y.},
  title     = {MultiMolecule},
  doi       = {10.5281/zenodo.12638419},
  publisher = {Zenodo},
  url       = {https://doi.org/10.5281/zenodo.12638419},
  year      = 2024,
  month     = may,
  day       = 4
}

Contact

Please use GitHub issues of MultiMolecule for any questions or comments on the model card.

Please contact the authors of the ERNIE-RNA paper for questions or comments on the paper/model.

License

This model implementation is licensed under the GNU Affero General Public License.

For additional terms and clarifications, please refer to our License FAQ.

SPDX-License-Identifier: AGPL-3.0-or-later

multimolecule/ernierna-ss

ERNIE-RNA

Disclaimer

Model Details

Variants

Model Specification

Links

Usage

Direct Use

RNA Secondary Structure Prediction

Downstream Use

Extract Features

Sequence Classification / Regression

Token Classification / Regression

Contact Classification / Regression

Training Details

Training Data

Training Procedure

Preprocessing

Pre-training

Citation

Contact

License

multimolecule/
ernierna-ss