Rnaernie

🚀 RNAErnie

A pre-trained model on non-coding RNA (ncRNA) using a multi-stage masked language modeling (MLM) objective.

🚀 Quick Start

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

pip install multimolecule

✨ Features

• Self-supervised Pretraining: RNAErnie is a bert-style model pre-trained on a large corpus of non-coding RNA sequences in a self-supervised fashion.
• Multi - stage Training: It uses a special 3 - stage training pipeline with different masking strategies.
• Multiple Downstream Tasks: Can be used for various downstream tasks such as sequence classification, token classification, and contact classification after fine - tuning.

📦 Installation

The installation of the necessary library can be done using the following command:

pip install multimolecule

💻 Usage Examples

Basic Usage

You can use this model directly with a pipeline for masked language modeling:

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

>>> unmasker = pipeline("fill-mask", model="multimolecule/rnaernie")
>>> unmasker("gguc<mask>cucugguuagaccagaucugagccu")
[{'score': 0.09252794831991196,
  'token': 8,
  'token_str': 'G',
  'sequence': 'G G U C G C U C U G G U U A G A C C A G A U C U G A G C C U'},
 {'score': 0.09062391519546509,
  'token': 11,
  'token_str': 'R',
  'sequence': 'G G U C R C U C U G G U U A G A C C A G A U C U G A G C C U'},
 {'score': 0.08875908702611923,
  'token': 6,
  'token_str': 'A',
  'sequence': 'G G U C A C U C U G G U U A G A C C A G A U C U G A G C C U'},
 {'score': 0.07809742540121078,
  'token': 20,
  'token_str': 'V',
  'sequence': 'G G U C V C U C U G G U U A G A C C A G A U C U G A G C C U'},
 {'score': 0.07325706630945206,
  'token': 13,
  'token_str': 'S',
  'sequence': 'G G U C S C U C U G G U U A G A C C A G A U C U G A G C C U'}]

Advanced Usage

Extract Features

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

from multimolecule import RnaTokenizer, RnaErnieModel


tokenizer = RnaTokenizer.from_pretrained("multimolecule/rnaernie")
model = RnaErnieModel.from_pretrained("multimolecule/rnaernie")

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

output = model(**input)

Sequence Classification / Regression

⚠️ Important 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.

import torch
from multimolecule import RnaTokenizer, RnaErnieForSequencePrediction


tokenizer = RnaTokenizer.from_pretrained("multimolecule/rnaernie")
model = RnaErnieForSequencePrediction.from_pretrained("multimolecule/rnaernie")

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

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

Token Classification / Regression

⚠️ Important 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.

import torch
from multimolecule import RnaTokenizer, RnaErnieForTokenPrediction


tokenizer = RnaTokenizer.from_pretrained("multimolecule/rnaernie")
model = RnaErnieForTokenPrediction.from_pretrained("multimolecule/rnaernie")

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

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

Contact Classification / Regression

⚠️ Important 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.

import torch
from multimolecule import RnaTokenizer, RnaErnieForContactPrediction


tokenizer = RnaTokenizer.from_pretrained("multimolecule/rnaernie")
model = RnaErnieForContactPrediction.from_pretrained("multimolecule/rnaernie")

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

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

📚 Documentation

Statement

Multi - purpose RNA language modelling with motif - aware pretraining and type - guided fine - tuning is published in Nature Machine Intelligence, which is a Closed Access / Author - Fee journal.

Machine learning has been at the forefront of the movement for free and open access to research.

We see no role for closed access or author - fee publication in the future of machine learning research and believe the adoption of these journals as an outlet of record for the machine learning community would be a retrograde step.

The MultiMolecule team is committed to the principles of open access and open science.

We do NOT endorse the publication of manuscripts in Closed Access / Author - Fee journals and encourage the community to support Open Access journals and conferences.

Please consider signing the Statement on Nature Machine Intelligence.

Disclaimer

This is an UNOFFICIAL implementation of the RNAErnie: An RNA Language Model with Structure - enhanced Representations by Ning Wang, Jiang Bian, Haoyi Xiong, et al.

The OFFICIAL repository of RNAErnie is at CatIIIIIIII/RNAErnie.

⚠️ Important Note

The MultiMolecule team is unable to confirm that the provided model and checkpoints are producing the same intermediate representations as the original implementation. This is because

The proposed method is published in a Closed Access / Author - Fee journal.

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

Model Details

RNAErnie 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.

Note that during the conversion process, additional tokens such as [IND] and ncRNA class symbols are removed.

Model Specification

Links

• Code: multimolecule.rnaernie
• Weights: multimolecule/rnaernie
• Data: multimolecule/rnacentral
• Paper: Multi - purpose RNA language modelling with motif - aware pretraining and type - guided fine - tuning
• Developed by: Ning Wang, Jiang Bian, Yuchen Li, Xuhong Li, Shahid Mumtaz, Linghe Kong, Haoyi Xiong.
• Model type: BERT - ERNIE
• Original Repository: CatIIIIIIII/RNAErnie

Training Details

RNAErnie 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 RNAErnie 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](https://rnacentral.org/expert - databases) representing a broad range of organisms and RNA types.

RNAErnie used a subset of RNAcentral for pre - training. The subset contains 23 million sequences. RNAErnie 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

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

• 15% of the tokens are masked.
• In 80% of the cases, the masked tokens are replaced by <mask>.
• In 10% of the cases, the masked tokens are replaced by a random token (different) from the one they replace.
• In the 10% remaining cases, the masked tokens are left as is.

Pre - training

RNAErnie used a special 3 - stage training pipeline to pre - train the model, each with a different masking strategy: Base - level Masking: The masking applies to each nucleotide in the sequence. Subsequence - level Masking: The masking applies to subsequences of 4 - 8bp in the sequence. Motif - level Masking: The model is trained on motif datasets.

The model was trained on 4 NVIDIA V100 GPUs with 32GiB memories.

• Batch size: 50
• Steps: 2,580,000
• Optimizer: AdamW
• Learning rate: 1e - 4
• Learning rate warm - up: 129,000 steps
• Learning rate cool - down: 129,000 steps
• Minimum learning rate: 5e - 5
• Weight decay: 0.01

Citation

Citation information is not available for papers published in Closed Access / Author - Fee journals.

Contact

Please use GitHub issues of [MultiMolecule](https://github.com/DLS5 - Omics/multimolecule/issues) for any questions or comments on the model card.

Please contact the authors of the RNAErnie paper for questions or comments on the paper/model.

📄 License

This model is licensed under the [AGPL - 3.0 License](https://www.gnu.org/licenses/agpl - 3.0.html).

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