GB.RNA-1.6B-CDS

GB.RNA-1.6B-CDS is a domain adaptation model on the coding sequences. It was pre-trained on 9 million coding sequences released by Carlos et al. (2024) [1] based on our GB.RNA-1.6B model.

How to Use

Build any downstream models from this backbone with ModelGenerator

For more information, visit: Model Generator

mgen fit --model SequenceClassification --model.backbone aido_rna_1b600m_cds --data SequenceClassificationDataModule --data.path <hf_or_local_path_to_your_dataset>
mgen test --model SequenceClassification --model.backbone aido_rna_1b600m_cds --data SequenceClassificationDataModule --data.path <hf_or_local_path_to_your_dataset>

Or use directly in Python

Embedding

from modelgenerator.tasks import Embed
model = Embed.from_config({"model.backbone": "aido_rna_1b600m_cds"}).eval()
transformed_batch = model.transform({"sequences": ["ACGT", "AGCT"]})
embedding = model(transformed_batch)
print(embedding.shape)
print(embedding)

Sequence-level Classification

import torch
from modelgenerator.tasks import SequenceClassification
model = SequenceClassification.from_config({"model.backbone": "aido_rna_1b600m_cds", "model.n_classes": 2}).eval()
transformed_batch = model.transform({"sequences": ["ACGT", "AGCT"]})
logits = model(transformed_batch)
print(logits)
print(torch.argmax(logits, dim=-1))

Token-level Classification

import torch
from modelgenerator.tasks import TokenClassification
model = TokenClassification.from_config({"model.backbone": "aido_rna_1b600m_cds", "model.n_classes": 3}).eval()
transformed_batch = model.transform({"sequences": ["ACGT", "AGCT"]})
logits = model(transformed_batch)
print(logits)
print(torch.argmax(logits, dim=-1))

Sequence-level Regression

from modelgenerator.tasks import SequenceRegression
model = SequenceRegression.from_config({"model.backbone": "aido_rna_1b600m_cds"}).eval()
transformed_batch = model.transform({"sequences": ["ACGT", "AGCT"]})
logits = model(transformed_batch)
print(logits)

Get RNA sequence embedding

from genbio_finetune.tasks import Embed
model = Embed.from_config({"model.backbone": "aido_rna_1b600m_cds"}).eval()
transformed_batch = model.transform({"sequences": ["ACGT", "ACGT"]})
embedding = model(transformed_batch)
print(embedding.shape)
print(embedding)

Citation

Please cite GB.RNA using the following BibTeX code:

@inproceedings{zou_large-scale_2024,
    title = {A Large-Scale Foundation Model for RNA Function and Structure Prediction},
    url = {https://www.biorxiv.org/content/10.1101/2024.11.28.625345v1},
    doi = {10.1101/2024.11.28.625345},
    publisher = {bioRxiv},
    author = {Zou, Shuxian and Tao, Tianhua and Mahbub, Sazan and Ellington, Caleb N. and Algayres, Robin and Li, Dian and Zhuang, Yonghao and Wang, Hongyi and Song, Le and Xing, Eric P.},
    year = {2024},
    booktitle = {NeurIPS 2024 Workshop on AI for New Drug Modalities},
}

Reference

1. Carlos Outeiral and Charlotte M Deane. Codon language embeddings provide strong signals for use in protein engineering. Nature Machine Intelligence, 6(2):170–179, 2024.