JUMP-lite: Compact, reproducible benchmarking of cell representations
Abstract
Image-based profiling captures rich phenotypic signatures for drug discovery and functional genomics. Large public datasets like JUMP Cell Painting now provide millions of images for systematic study. However, JUMP alone occupies 115 TB, and fragmented evaluation practices make systematic comparisons of representation methods impractical for many researchers. Here we present Nahual, an open-source framework for reproducible model deployment, and JUMP-lite, a 92.0 GB subset of JUMP that is approximately 1,250-fold smaller, selected to cover genetic modalities and compound annotations and reduced via lossy JPEG XL compression. Using these resources, we benchmark five representation methods, including classical features (CellProfiler) and deep learning models (MorphEM, OpenPhenom, SubCell, DINOv2). Moderate compression broadly retains signal relative to uncompressed images. Standardized phenotypic activity and consistency metrics reveal meaningful performance differences across methods. Together, JUMP-lite and Nahual provide a foundation for accessible, reproducible benchmarking of image-based cell representations.
Related
- A mechanism-guided simulation framework for synthetic tax-risk data: Constraint validation and reproducible benchmarking
- Distinct genetic requirements underlie LITE-1-dependent photosensation and avoidance of high concentration of diacetyl in C. elegans
- A highly reproducible, safe cranial window surgery for targeted in vivo manipulations in the adult zebrafish telencephalon
- Cross-Dataset Reanalysis of Mouse Alzheimer Disease Single-Cell RNA-Sequencing Data Reveals a Reproducible Innate Immune Activation Signature
- OmicsResonance: An LLM-assisted cloud ecosystem for interactiveand reproducible single-cell transcriptomic analysis
BibTeX
@misc{munoz2026jumplite,
title = {{JUMP-lite: Compact, reproducible benchmarking of cell representations}},
author = {Alán F. Muñoz and Johan Fredin Haslum and Runxi Shen and Anne E. Carpenter and Shantanu Singh},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2608.07632},
eprint = {2608.07632},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2608.07632},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - JUMP-lite: Compact, reproducible benchmarking of cell representations
AU - Alán F. Muñoz
AU - Johan Fredin Haslum
AU - Runxi Shen
AU - Anne E. Carpenter
AU - Shantanu Singh
PY - 2026
DA - 2026/09/01
PB - arXiv
DO - 10.48550/arxiv.2608.07632
UR - https://arxiv.org/abs/2608.07632
AB - Image-based profiling captures rich phenotypic signatures for drug discovery and functional genomics. Large public datasets like JUMP Cell Painting now provide millions of images for systematic study. However, JUMP alone occupies 115 TB, and fragmented evaluation practices make systematic comparisons of representation methods impractical for many researchers. Here we present Nahual, an open-source framework for reproducible model deployment, and JUMP-lite, a 92.0 GB subset of JUMP that is approximately 1,250-fold smaller, selected to cover genetic modalities and compound annotations and reduced via lossy JPEG XL compression. Using these resources, we benchmark five representation methods, including classical features (CellProfiler) and deep learning models (MorphEM, OpenPhenom, SubCell, DINOv2). Moderate compression broadly retains signal relative to uncompressed images. Standardized phenotypic activity and consistency metrics reveal meaningful performance differences across methods. Together, JUMP-lite and Nahual provide a foundation for accessible, reproducible benchmarking of image-based cell representations.
ER -