Cell type proportion estimates from high quality images from Huuki-Myers et al., bioRxiv, 2024, doi: https://doi.org/10.1101/2024.02.09.579665.
data("RNAScope_prop")A data.frame object.
https://github.com/LieberInstitute/brainstorm/blob/master/inst/extdata/data-raw/RNAScope_prop.R
11.49 kB
These are the columns of the data.frame object:
SAMPLE_ID : DLPFC Tissue block + RNAScope combination.
Sample : DLFPC Tissue block (Donor BrNum + DLPFC position).
Combo : RNAScope probe combination, either "Circle" marking cell types Astro
Endo, Inhib, or "Star" marking Excit, Micro, and OligoOPC.
cell_type : The cell type measured.
Confidence : Image confidence, this dataset has been filtered to the high & Ok confidence images.
n_cell : the number of cells counted for the Sample and cell type.
prop : the calculated cell type proportion from n_cell.
n_cell_sn : number of nuclei in the corresponding snRNA-seq data.
prop_sn : cell type proportion from the snRNA-seq data.
NOTE: For the RNAScope assay utilized here only 3 cell types could be measured at once. Two consecutive tissue slices were used from each brain block to measure one combination of three major cell types, then the other three (differentiated as circle and star). DAPI was used to mark the nuclei, in each tissue section, so the overall number of cells is recorded but only a fraction has a cell type label, unlabeled nuclei are classified "other".
The two sections combined should get close to identifying the type of all the cells, but often the combined "non-other" fractions are around 0.7 and not 1. This could be due to a few reasons: the sections while as similar as possible are not the same tissue, error in the assay and/or image processing not labeling all cells possible, or presence of rare cell types not marked by the selected probes.
With all of this considered, we still think the RNAScope estimates are good approximations of the cell type proportions in these samples.
For more info check out https://doi.org/10.1101/2024.02.09.579665 Figure 2, and Methods: RNAScope/Immunofluorescence Data Generation and HALO Analysis.
# To view source
system.file("extdata", "data-raw", "RNAScope_prop.R", package = "DeconvoBuddies")
#> [1] "/__w/_temp/Library/DeconvoBuddies/extdata/data-raw/RNAScope_prop.R"