Datasets:

cp-bg-bench-anon
/

jump-sample

license: cc-by-4.0
language:
  - en
pretty_name: cp-bg-bench preview (jump)
size_categories:
  - 1K<n<10K
task_categories:
  - image-classification
tags:
  - cell-painting
  - microscopy
  - biology
  - benchmark
configs:
  - config_name: crops
    data_files:
      - split: train
        path: crops/data-*.arrow
  - config_name: crops_density
    data_files:
      - split: train
        path: crops_density/data-*.arrow
  - config_name: seg
    data_files:
      - split: train
        path: seg/data-*.arrow
  - config_name: seg_density
    data_files:
      - split: train
        path: seg_density/data-*.arrow

cp-bg-bench preview — jump

Compact preview of the jump dataset from the cp-bg-bench benchmark. Stratified subset of the full release; designed so that the held-out-batch perturbation-recall and cp_measure-prediction evals can be reproduced end-to-end against this small slice alone.


Cells	913
Wells	69
Perturbations	33
Held-out batch	`source_4`
Views	crops, crops_density, seg, seg_density

What's in this repo

crops/                # HF dataset, 224×224×C uint8 cell crops + masks
crops_density/        # crops + 4 corner density patches
seg/                  # mask-applied cell crops
seg_density/          # mask + density patches
aggregated/           # well-level aggregated h5ad with X_pca + X_pca_harmony
cp_measure/           # well-level cp_measure features (~1300-1400 features)
metadata/             # perturbations parquet + quality filter report
build_info.json       # full provenance (seed, source pipeline commit, selection stats)

Schema (per cell row)

row_key, source, plate, well, tile, id_local                              # identity
nuc_area, cyto_area, nuc_cyto_ratio, n_cells_in_fov, n_cells_scaled        # per-cell QC
mask:  large_binary  # (2, 224, 224) uint8 -> NucMask, CellMask
cell:  large_binary  # (C, 224, 224) uint8 fluorescence channels
Metadata_JCP2022, Metadata_InChIKey, Metadata_PlateType                    # source metadata
perturbation, batch, treatment, Metadata_Perturbation                      # benchmark axes

Selection recipe

All control wells, capped to a small budget.
Treated wells whose perturbation appears in both the held-out batch and at least one non-held-out batch (so cross-batch retrieval is reproducible). Per chosen perturbation: one held-out well + one non-held-out well.
Per-well cell cap to keep total budget near 1,000 cells.

Seed = 42. See build_info.json for exact stats.

Quick start

from datasets import load_dataset

ds = load_dataset("cp-bg-bench-anon/jump-sample", name="crops", split="train")
print(ds)
print(ds[0]["row_key"], ds[0]["perturbation"])

Full dataset

The corresponding full-size release is hosted separately.

License

JUMP-CP: CC-BY 4.0. RxRx1 / RxRx3-core: CC BY-NC-SA 4.0. Verify upstream license terms before use.