Segmentation for methylation noise reduction: input data bundle (RRBS methylation matrices and ChromHMM annotations)

Input data bundle for the Segmentation for methylation noise reduction project. This archive contains all external data files required to run project notebooks end-to-end. CONTENTS Methylation data (three studies): Petkovich_GSE80672_Efi_processed/ — preprocessed mouse blood RRBS as parquet beta-value matrices (CpG × sample, NaN where coverage ≤ 5 reads) with metadata CSVs. Used in notebook 01 for training epigenetic age clocks. Source: Petkovich et al., 2017 (Cell Metabolism, doi:10.1016/j.cmet.2017.03.016), GEO GSE80672. Preprocessing by Evgeniy Efimov. Thompson_GSE120132_Efi_processed/ — preprocessed mouse multi-tissue RRBS, blood subset, in the same parquet format. Used in notebook 01 for external clock validation. Source: Thompson et al., 2018 (Aging, doi:10.18632/aging.101590), GEO GSE120132. Preprocessing by Evgeniy Efimov. obesity_GSE85928/ — raw bigWig methylation files from human whole blood (lean vs obese), as deposited on GEO. Used in notebook 02 for the group separation benchmark. Source: Day et al., 2017 (Epigenetics, doi:10.1080/15592294.2017.1281501GEO GSE85928. No further preprocessing applied. ChromHMM chromatin-state annotations: mouse_full_stack_ChromHMM_annotations/ — mouse full-stack ChromHMM annotation (mm10, 100 states). Source: Vu & Ernst, 2023 (Genome Biology, doi:10.1186/s13059-023-02994-x). Original files distributed via https://public.hoffman2.idre.ucla.edu/ernst/2K9RS//full_stack/full_stack_annotation_public_release/mm10/mm10_100_segments_segments.bed.gz human_Roadmap_Epigenomics_chromhmmSegmentations/ — human ChromHMM segmentations from the Roadmap Epigenomics project (15-state core model, E062 PBMC). Source: Roadmap Epigenomics Consortium, 2015 (Nature, doi:10.1038/nature14248). Original files distributed via https://egg2.wustl.edu/roadmap/data/byFileType/chromhmmSegmentations/ChmmModels/coreMarks/jointModel/final/E062_15_coreMarks_dense.bed.gz. USAGE Download and extract this archive so that the contents form a sources/ folder at the root of the repository. Then run the notebooks top-to-bottom — they will find all required inputs at the expected paths. LICENSING The preprocessed parquet matrices (Petkovich and Thompson) are released under CC BY 4.0 with attribution to Evgeniy Efimov. Raw bigWig files in obesity_GSE85928/ retain the original GEO terms of use. ChromHMM annotation files are bundled verbatim from their public distribution points and retain their original licensing. When using any component, please cite the original publications listed above.