suke-sho's picture
Update README.md
f650cde verified
---
license: apache-2.0
tags:
- biology
- dna
---
# Dataset Card for Arabidopsis Genome Corpus
***
### Dataset Description
This dataset consists of model plant genomes available on NCBI, specifically processed and cleaned for research and pre-training purposes.
The default configuration, "6kbp," provides DNA chunks of 6.2kbp with a 100bp overlap on each side.
The sequences are cleaned and processed to contain only the letters A, T, C, G, and N, ensuring high-quality genomic data suitable for various bioinformatics and machine learning applications.
### Dataset Summary
- Homepage: NCBI
- License: NCBI Policies
- Version: 1.1.0
- Size: Variable depending on the configuration
- Default Configuration: 6kbp chunks
### Usage Example
```
from datasets import load_dataset
# default chunk-size is 6000
chunk_length = 1024
repo = "suke-sho/arabidopsis_genome_corpus"
datasets = load_dataset(repo, trust_remote_code=True, chunk_length=chunk_length)
```
### Acknowledgments
This dataset is an adaptation of the "InstaDeepAI/plant-multi-species-genomes" dataset & script.
Special thanks to InstaDeepAI for their original work.