Supporting Data for OmniAge Package

Supporting Data for OmniAge: Model Weights & Example Datasets This repository hosts the essential data dependencies for the OmniAge ecosystem. It includes pre-trained model weights for epigenetic clocks and example datasets for testing both the Python and R versions of the package. 📂 File Contents Model Weights (Core Data): PCClocks.tar.gz: Pre-trained weights converted to Parquet format (Primary dependency for omniage Python package). SystemsAge.tar.gz: Pre-trained weights converted to Parquet format. Python Examples: OmniAge_py_example.tar.gz: Sample input data and scripts for the Python package. R Examples (Individual Files): *.rda files: A collection of standalone example datasets for the R package (OmniAgeR). 🧬 Data Origin & Attribution (Model Weights) The model weights (PC Clocks & SystemsAge) are derived from the original data released by Albert Higgins-Chen et al. Original Dataset Title: PCClock and SystemsAge Data Original Author: Higgins-Chen, Albert Original DOI: 10.5281/zenodo.17162604 Original Repository: https://github.com/HigginsChenLab/methylCIPHER 🔄 Modifications to Model Weights The original model data was distributed in .qs2 (R format). To facilitate usage within the Python ecosystem, we performed the following conversions: Format Shift: Converted from R qs2 binary format to Apache Parquet (.parquet). Compression: Applied zstd compression. Integrity: The numerical values remain identical to the original source. 📚 References If you use these models, please cite both this repository AND the original papers: Original Data: Higgins-Chen, A. (2025). PCClock and SystemsAge Data [Data set]. Zenodo. https://doi.org/10.5281/zenodo.17162604 PC Clocks Paper: Higgins-Chen, A. T., et al. (2022). Nature Aging. SystemsAge Paper: Sehgal, R., et al. (2024). Nature Aging.