---
language:
- en
license: other
tags:
- biology
- proteins
- sequence-classification
- benchmark
task_categories:
- text-classification
pretty_name: SignalP_Binary
---

# SignalP_Binary

Binary benchmark for signal peptide prediction from protein sequences, adapted from the ProteinBERT benchmark collection and SignalP data.

## Source

This dataset is sourced from the ProteinBERT benchmark repository:
https://github.com/nadavbra/protein_bert/tree/master/protein_benchmarks

## Curator Attribution

This Hugging Face dataset packaging, curation, and publication was prepared by Dan Ofer.

## Splits and Schema

- Splits follow the benchmark release (train/validation/test when available).
- Each row includes:
  - seq: amino-acid sequence
  - label: binary target (0 or 1)

## Hugging Face Repo

- GrimSqueaker/SignalP_Binary

## Citations

```bibtex
@article{Armenteros2019,
    author = {Almagro Armenteros, Jose Juan and Tsirigos, Konstantinos D. and Sonderby, Casper Kaae and Petersen, Thomas Nordahl and Winther, Ole and Brunak, Soren and von Heijne, Gunnar and Nielsen, Henrik},
    title = {SignalP 5.0 improves signal peptide predictions using deep neural networks},
    journal = {Nature Biotechnology},
    volume = {37},
    number = {4},
    pages = {420--423},
    year = {2019},
    doi = {10.1038/s41587-019-0036-z}
}

@article{10.1093/bioinformatics/btac020,
    author = {Brandes, Nadav and Ofer, Dan and Peleg, Yam and Rappoport, Nadav and Linial, Michal},
    title = {ProteinBERT: a universal deep-learning model of protein sequence and function},
    journal = {Bioinformatics},
    volume = {38},
    number = {8},
    pages = {2102-2110},
    year = {2022},
    doi = {10.1093/bioinformatics/btac020}
}
```