README.md

# Dataset for essential genes prediction in bacterial genomes (DNA)

A dataset of 169,408 genes with gene essentiality labels (binary) from 51 bacterial genomes across 37 species.

The gene essentiality labels have been extracted from the [Database of Essential Genes](http://origin.tubic.org/deg/public/index.php/browse/bacteria) and the genome DNA sequences have been extracted from [GenBank](https://www.ncbi.nlm.nih.gov/genbank/). 
Each row contains whole bacterial genome together with annotated gene locations on the genome (`start` and `end`). We excluded duplicates and genomes with incomplete genomes.

## Labels
The labels for each gene are provided in the dataset in the `essential` column.

## 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-essential-genes-dna", split="validation")

# fetch an example
item = next(iter(ds))
# get gene sequences
gene_dna_seqs = [
  i["sequence"][int(start - 1):int(end)] for start, end in zip(item["start"], item["end"])
]
# get labels
labels = item["essential"]
```

## Split

We provide `60 / 20 /20` (%) `train`, `validation` and `test` splits respectively as part of the dataset. The split has been performed by randomly.

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 gene essentiality. For coding sequence representation of the genome see the [macwiatrak/bacbench-essential-genes-protein-sequences](https://huggingface.co/datasets/macwiatrak/bacbench-essential-genes-protein-sequences)
dataset.


---
dataset_info:
  features:
  - name: genome_name
    dtype: string
  - name: essential
    sequence: string
  - name: start
    sequence: int64
  - name: end
    sequence: int64
  - name: protein_id
    sequence: string
  - name: strand
    sequence: float64
  - name: product
    sequence: string
  - name: dna_seq
    dtype: string
  - name: __index_level_0__
    dtype: int64
  splits:
  - name: train
    num_bytes: 120955501
    num_examples: 30
  - name: validation
    num_bytes: 45990995
    num_examples: 11
  - name: test
    num_bytes: 36170463
    num_examples: 10
  download_size: 92709156
  dataset_size: 203116959
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
  - split: test
    path: data/test-*
---