BCTMA dataset
Description
Dataset associated with manuscript titled "Toward reproducible, scalable, and robust data analysis across multiplex tissue imaging platforms".
Descriptions of column names are below.
source: name of TMA
scene: unique core id for particular TMA
cell: unique cell id for particular source and scene
centroid_<x,y>: physical location of cell (0.325um/pixel)
<marker>_<compartment>: mean intensity for the given compartment, e.g. pHH3_Nuclei, Ecad_Ring. Here "Ring" means mean intensity from area defined by the full cell segmentation mask minus the nuclear segmentation mask, i.e. a 2D approximation of the cytoplasm.
Enorm_scene_<marker>_<compartment>: Same as above, but used RESTORE normalization for each core for each marker.
40_neighbor: PhenoGraph cluster label (we set k=40 when building the k-nearest neighbor graph)
metacluster: aggregated PhenoGraph cluster labels, based on hierarchical clustering to group cells in similar non-functional epigenetic/proliferative/stress states and to preserve known biological differences.
metacluster_name: semantic label for metacluster based on marker expression.
Files
final_metacluster_df.csv
Files
(1.9 GB)
| Name | Size | Download all |
|---|---|---|
|
md5:729b68b4acde03ccfe99ac353b5493c9
|
1.9 GB | Preview Download |