Published September 19, 2025 | Version v1
Dataset Open

Aldolase Test Data for prismPYP: Power-spectrum and image domain learning for self-supervised micrograph evaluation

  • 1. ROR icon Duke University

Description

This dataset contains training data, model weights, and sample selection results for performing self-supervised micrograph evaluation on the data deposited in EMPIAR-10379: "Cryo-EM motion corrected micrographs of rabbit muscle aldolase" by Li et al. (2020).
  • model_weights.tar.gz: Trained real domain and Fourier domain model weights
  • example_data.tar.gz:
    • .webp image files of micrographs and power spectra
    • .pkl metadata for the micrographs, obtained after processing the micrographs in nextPYP [Liu et al., 2023]
    • J7_exposures_accepted_exported.cs file containing metadata from cryoSPARC [Punjani et al., 2017]
    • sp-preprocessing-fhgRaEnEqUsEFrUj.micrographs file containing a list of the images present in the dataset
    • .pyp_config.toml file containing microscope parameters for this data collection session
  • fft_good_export.parquet: Data points that have high-quality features in the Fourier domain
  • real_good_export.parquet: Data points that have high-quality features in the real domain

Files

Files (1.4 GB)

Name Size Download all
md5:f4bdf267883784c360a8011948d6b9d2
1.2 GB Download
md5:7ef4554f851d6e6296b4d210e7fbecd9
3.2 MB Download
md5:fa11d99c0c3acd43560cb7235431f484
233.7 MB Download
md5:45d33be6a4b47ac37e19de13249ff5f7
3.1 MB Download