README.md

---
language: dna
tags:
  - Biology
  - DNA
  - Genomics
license: agpl-3.0
datasets:
  - multimolecule/gencode
library_name: multimolecule
---

# OpenSpliceAI

Modular native-PyTorch reimplementation of SpliceAI for predicting pre-mRNA splice sites from primary DNA sequence.

## Disclaimer

This is an UNOFFICIAL implementation of [OpenSpliceAI: An efficient, modular implementation of SpliceAI enabling easy retraining on non-human species](https://doi.org/10.7554/eLife.107454.3) by Kuan-Hao Chao, Alan Mao, Anqi Liu, Steven L. Salzberg and Mihaela Pertea.

The OFFICIAL repository of OpenSpliceAI is at [Kuanhao-Chao/OpenSpliceAI](https://github.com/Kuanhao-Chao/OpenSpliceAI).

> [!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 OpenSpliceAI did not write this model card for this model so this model card has been written by the MultiMolecule team.**

## Model Details

OpenSpliceAI is a deep dilated residual convolutional neural network that reimplements the SpliceAI architecture in native PyTorch. It predicts, for each nucleotide of a pre-mRNA transcript, whether the position is a splice acceptor, a splice donor, or neither. The model stacks dilated residual units with increasing kernel size and atrous rate so that a wide genomic context window contributes to each per-nucleotide prediction, while skip connections aggregate multi-scale features. OpenSpliceAI reproduces the predictive behavior of SpliceAI while providing an efficient, modular training pipeline that can be retrained on non-human species.

### Variations

OpenSpliceAI ships trained model families for human MANE and four non-human species. Each family provides four
flanking-context sizes. MultiMolecule publishes one seed (`rs10`) for each family/context pair; the other upstream seed
checkpoints are training replicates and are not exposed as separate model variants.

| Family        | 80 nt                                                                                                 | 400 nt                                                                                                  | 2,000 nt                                                                                                  | 10,000 nt                                                                                                   |
| ------------- | ----------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------- |
| MANE / human  | [`openspliceai-mane-80nt`](https://huggingface.co/multimolecule/openspliceai-mane-80nt)               | [`openspliceai-mane-400nt`](https://huggingface.co/multimolecule/openspliceai-mane-400nt)               | [`openspliceai-mane-2000nt`](https://huggingface.co/multimolecule/openspliceai-mane-2000nt)               | [`openspliceai-mane-10000nt`](https://huggingface.co/multimolecule/openspliceai-mane-10000nt)               |
| Mouse         | [`openspliceai-mouse-80nt`](https://huggingface.co/multimolecule/openspliceai-mouse-80nt)             | [`openspliceai-mouse-400nt`](https://huggingface.co/multimolecule/openspliceai-mouse-400nt)             | [`openspliceai-mouse-2000nt`](https://huggingface.co/multimolecule/openspliceai-mouse-2000nt)             | [`openspliceai-mouse-10000nt`](https://huggingface.co/multimolecule/openspliceai-mouse-10000nt)             |
| Zebrafish     | [`openspliceai-zebrafish-80nt`](https://huggingface.co/multimolecule/openspliceai-zebrafish-80nt)     | [`openspliceai-zebrafish-400nt`](https://huggingface.co/multimolecule/openspliceai-zebrafish-400nt)     | [`openspliceai-zebrafish-2000nt`](https://huggingface.co/multimolecule/openspliceai-zebrafish-2000nt)     | [`openspliceai-zebrafish-10000nt`](https://huggingface.co/multimolecule/openspliceai-zebrafish-10000nt)     |
| Honeybee      | [`openspliceai-honeybee-80nt`](https://huggingface.co/multimolecule/openspliceai-honeybee-80nt)       | [`openspliceai-honeybee-400nt`](https://huggingface.co/multimolecule/openspliceai-honeybee-400nt)       | [`openspliceai-honeybee-2000nt`](https://huggingface.co/multimolecule/openspliceai-honeybee-2000nt)       | [`openspliceai-honeybee-10000nt`](https://huggingface.co/multimolecule/openspliceai-honeybee-10000nt)       |
| _Arabidopsis_ | [`openspliceai-arabidopsis-80nt`](https://huggingface.co/multimolecule/openspliceai-arabidopsis-80nt) | [`openspliceai-arabidopsis-400nt`](https://huggingface.co/multimolecule/openspliceai-arabidopsis-400nt) | [`openspliceai-arabidopsis-2000nt`](https://huggingface.co/multimolecule/openspliceai-arabidopsis-2000nt) | [`openspliceai-arabidopsis-10000nt`](https://huggingface.co/multimolecule/openspliceai-arabidopsis-10000nt) |

### Model Specification

| Flanking Context | Residual Blocks | Hidden Size | Num Parameters (M) | FLOPs (G) | MACs (G) |
| ---------------- | --------------- | ----------- | ------------------ | --------- | -------- |
| 80 nt            | 4               | 32          | 0.09               | 0.95      | 0.47     |
| 400 nt           | 8               | 32          | 0.19               | 2.00      | 0.99     |
| 2,000 nt         | 12              | 32          | 0.36               | 5.03      | 2.50     |
| 10,000 nt        | 16              | 32          | 0.70               | 20.90     | 10.40    |

Model size is determined by flanking context and is shared across species for the same context. FLOPs and MACs are
reported for a single 5,000-nucleotide output sequence.

### Links

- **Code**: [multimolecule.openspliceai](https://github.com/DLS5-Omics/multimolecule/tree/master/multimolecule/models/openspliceai)
- **Weights**: See the 20 variant repositories listed above.
- **Data**: Human MANE/GENCODE for the MANE variants; species annotations follow the original OpenSpliceAI release.
- **Paper**: [OpenSpliceAI: An efficient, modular implementation of SpliceAI enabling easy retraining on non-human species](https://doi.org/10.7554/eLife.107454.3)
- **Developed by**: Kuan-Hao Chao, Alan Mao, Anqi Liu, Steven L. Salzberg, Mihaela Pertea
- **Original Repository**: [Kuanhao-Chao/OpenSpliceAI](https://github.com/Kuanhao-Chao/OpenSpliceAI) (licensed GPL-3.0)

## Usage

The model file depends on the [`multimolecule`](https://multimolecule.danling.org) library. You can install it using pip:

```bash
pip install multimolecule
```

### Direct Use

#### RNA Splicing Site Prediction

You can use this model directly to predict the splice sites of a pre-mRNA sequence:

```python
>>> from multimolecule import DnaTokenizer, OpenSpliceAiForTokenPrediction

>>> model_id = "multimolecule/openspliceai-mouse-2000nt"
>>> tokenizer = DnaTokenizer.from_pretrained(model_id)
>>> model = OpenSpliceAiForTokenPrediction.from_pretrained(model_id)
>>> output = model(tokenizer("AGCAGTCATTATGGCGAA", return_tensors="pt")["input_ids"])

>>> output.keys()
odict_keys(['logits'])
```

Each output position carries three logits corresponding to _neither_, _acceptor_, and _donor_.

## Training Details

OpenSpliceAI was trained to predict the location of splice donor and acceptor sites from primary DNA sequence, following the SpliceAI training methodology.

### Training Data

The MANE variants were trained on transcripts from the [GENCODE](https://multimolecule.danling.org/datasets/gencode)/MANE human reference annotation. The non-human variants use the species annotations released by OpenSpliceAI for mouse, zebrafish, honeybee, and _Arabidopsis_. For each predicted nucleotide, the model receives a flanking context of 80, 400, 2,000, or 10,000 nucleotides, split evenly across the two sides of the output sequence, with sequence ends padded with `N`. Annotated splice donor and acceptor sites serve as positive labels; all other positions are negative.

### Training Procedure

#### Pre-training

The model was trained to minimize a cross-entropy loss between predicted splice-site probabilities and the reference annotation.

- Architecture: dilated residual 1D convolutions with skip connections
- Activation: LeakyReLU (slope 0.1)
- Optimizer: Adam
- Loss: cross-entropy
- Flanking context sizes: 80 / 400 / 2000 / 10000 nucleotides

Please refer to the [OpenSpliceAI paper](https://doi.org/10.7554/eLife.107454.3) for the full training protocol and hardware details.

## Citation

```bibtex
@article{chao2025openspliceai,
  author    = {Chao, Kuan-Hao and Mao, Alan and Liu, Anqi and Salzberg, Steven L and Pertea, Mihaela},
  title     = {OpenSpliceAI: An efficient, modular implementation of SpliceAI enabling easy retraining on non-human species},
  journal   = {eLife},
  volume    = 14,
  pages     = {RP107454},
  year      = 2025,
  doi       = {10.7554/eLife.107454.3},
  publisher = {eLife Sciences Publications, Ltd}
}
```

> [!NOTE]
> The artifacts distributed in this repository are part of the MultiMolecule project.
> If you use MultiMolecule in your research, you must 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](https://github.com/DLS5-Omics/multimolecule/issues) for any questions or comments on the model card.

Please contact the authors of the [OpenSpliceAI paper](https://doi.org/10.7554/eLife.107454.3) for questions or comments on the paper/model.

## License

This model implementation is licensed under the [GNU Affero General Public License](license.md).

For additional terms and clarifications, please refer to our [License FAQ](license-faq.md).

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