---
license: apache-2.0
tags:
- ppi
- proteins
- biology
- bacteria
- stringdb
pretty_name: Dataset for predicting protein-protein interactions in bacterial genomes
size_categories:
- 10K<n<100K
---
# Dataset for protein-protein interaction prediction across bacteria (Protein sequences)

A dataset of 10,533 bacterial genomes across 6,956 species with protein-protein interaction (PPI) scores for each genome.

The genome protein sequences and PPI scores have been extracted from [STRING DB](https://string-db.org/). 
Each row contains a set of protein sequences from a genome, ordered by their location on the chromosome and plasmids and a set of associated PPI scores.
The PPI scores have been extracted using the `combined` score from STRING DB. 


The interaction between two proteins is represented by a triple: `[prot1_index, prot2_index, score]`. Where to get a probability score, you must divide the score by `1000` 
(i.e. if the score is `721` then to get a true score do `721/1000=0.721`). The index of a protein refers to the index of the protein in the `protein_sequences` column of the
row. See example below in [Usage](#usage)


## Usage
We recommend loading the dataset in a streaming mode to prevent memory errors.
```python
from datasets import load_dataset


ds = load_dataset("macwiatrak/bacbench-ppi-stringdb-protein-sequences", split="validation", streaming=True)
item = next(iter(ds))

# fetch protein sequences from a genome (list of strings)
prot_seqs = item["protein_sequences"]
# fetch PPI triples labels (i.e. [prot1_index, prot2_index, score])
ppi_triples = item["triples_combined_score"]

# get protein seqs and label for one pair of proteins
prot1 = prot_seqs[ppi_triples[0][0]]
prot2 = prot_seqs[ppi_triples[0][1]]
score = ppi_triples[0][2] / 1000

# we recommend binarizing the labels based on the threshold of 0.6
binary_ppi_triples = [
  (prot1_index, prot2_index, int((score / 1000) >= 0.6)) for prot1_index, prot2_index, score in ppi_triples
]
```

## Split

We provide `train`, `validation` and `test` splits with proportions of `70 / 10 / 20` (%) respectively as part of the dataset. The split was performed randomly at genome level.

See [github repository](https://github.com/macwiatrak/Bacbench) for details on how to embed the dataset with DNA and protein language models as well as code to predict antibiotic
resistance from sequence.

---
dataset_info:
  features:
  - name: taxid
    dtype: int64
  - name: genome_name
    dtype: string
  - name: protein_sequences
    sequence: string
  - name: protein_ids
    sequence: string
  - name: triples_combined_score
    sequence:
      sequence: int64
  - name: __index_level_0__
    dtype: int64
  splits:
  - name: train
    num_bytes: 169097180197
    num_examples: 7354
  - name: validation
    num_bytes: 27541966357
    num_examples: 1091
  - name: test
    num_bytes: 50704862883
    num_examples: 2088
  download_size: 53853247305
  dataset_size: 247344009437
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
  - split: test
    path: data/test-*
---