PLExBench: Maize and Arabidopsis Leaf and Root Gene Expression Benchmark for Sequence-to-Expression Models (Buckler Lab, Cornell)

Provider
Ferebee, Wrightsman, Zhai, Gokaslan, Kuleshov and Buckler (Cornell University)
License
MIT (source)
Size
4 files
Stored at
Hugging Face (https://huggingface.co/datasets/maize-genetics/plexbench-base)
Access
Free and public. No account with the data provider is needed.
Category
Genomics & Bioinformatics
Source
https://huggingface.co/datasets/maize-genetics/plexbench-base

About this dataset

Benchmark data for predicting plant gene expression from DNA sequence, from PLExBench: A benchmarking suite for predicting gene expression in plants, curated by Taylor Ferebee, Travis Wrightsman, Jingjing Zhai, Aaron Gokaslan, Volodymyr Kuleshov and Edward S. Buckler (Cornell University). Expression comes from public RNA-seq runs in the NCBI Sequence Read Archive for leaf and root tissue of Arabidopsis thaliana Col-0 and maize (Zea mays) B73 seedlings, quantified as transcripts per million (TPM). Three task types are derived per species and tissue: maximum expression across experiments, absolute (mean) expression, and on/off expression. Genes are split 80/10/10 into train, validation and test by orthogroup-guided splitting (Washburn et al. 2019) so related genes do not leak across splits. The repository holds two zip archives (about 0.8 GB): tasks/ with species-task-tissue folders of train, validate and test TSV files, and genomes/ with annotation FASTA and GFF files for each genome. Useful for training and evaluating DNA language models and sequence-to-expression predictors in crops. Released by the authors under the MIT License; cite the PLExBench paper.

How to load this dataset in Python

from huggingface_hub import hf_hub_download
path = hf_hub_download(repo_id="maize-genetics/plexbench-base", filename="genomes.zip", repo_type="dataset")
print(path)

Data structure

Files

Example file: genomes.zip

Analyze this dataset without downloading it

On Scigantic, PLExBench: Maize and Arabidopsis Leaf and Root Gene Expression Benchmark for Sequence-to-Expression Models (Buckler Lab, Cornell) opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.

This page is also available as Markdown and as JSON.

Related Genomics & Bioinformatics datasets