plantcad/vertebrate-genomes-plantcad2-c4096

---
license: apache-2.0
tags:
- biology
- DNA
- genomics
- genetics
- vertebrates
dataset_info:
features:
- name: text
dtype: string
splits:
- name: train
num_examples: 2638656
- name: validation
num_examples: 329832
- name: test
num_examples: 329832
---

# Vertebrate Genomes PlantCAD2 Subset (4096bp)

This dataset is a curated subset of [emarro/vertebrate_genomes](https://huggingface.co/datasets/emarro/vertebrate_genomes)
designed for comparative spectral analysis with plant genomic data.

## Dataset Description

Sequences were randomly sampled from Vertebrate Genomes (revision 9703952e2c90c822ea8a96c9638b584ccaf36d4e),
truncated to match the sample sizes per split of the
[plantcad/Angiosperm_65_genomes_8192bp](https://huggingface.co/datasets/plantcad/Angiosperm_65_genomes_8192bp) dataset.

### Processing Steps

1. **Streaming**: Records were streamed from the standard train/validation/test splits
2. **Shuffling**: Applied shuffle with buffer size of 10,000 for random sampling
3. **Truncation**: All sequences (originally 12kbp) were truncated to exactly 4096bp
4. **Sampling**: Collected samples to match PlantCAD split sizes

### Split Sizes

| Split | Number of Examples |
|-------|-------------------|
| train | 2,638,656 |
| validation | 329,832 |
| test | 329,832 |

### Sequence Length

All sequences are exactly **4096 base pairs**.

## Source Dataset

Vertebrate Genomes contains DNA sequences from vertebrate species, with all sequences being 12,000 base pairs in length.
This subset uses revision `9703952e2c90c822ea8a96c9638b584ccaf36d4e`.

## Usage

```python
from datasets import load_dataset

dataset = load_dataset("plantcad/vertebrate-genomes-plantcad2-c4096")
```

## License

Apache 2.0