spot- vs cell-level data access
1
0
Entering edit mode
kstern • 0
@1c1d1972
Last seen 5 weeks ago

Hi, thank you for your work!

I am a bit confused as to obtaining the spot vs cell level data. Following the spatialLIBD documentation, under the Processed Data section, to download spot-level data, one must call spe:

## Download the spot-level data
spe <- fetch_data(type = "spe")

## This is a SpatialExperiment object
spe


However, from a second documentation page, under the same section, fetching sce SingleCellExperiment data also claims to be spot-level data?

## Download the spot-level data
sce <- fetch_data(type = 'sce')
#>   sce

## This is a SingleCellExperiment object


I did notice that the SCE object is lacking spatialCoords but from the preprint, colData should "contain metadata describing the spatial coordinates" (or alternatively, "spatial coordinates of the cell centroids are stored in spatialData")....

I was hoping someone could clarify the spot-level (spe) vs. cell-level (sce) data and how to access them. I was also wondering, if the cell-level data is representative of the same spot-level data (sample) but at higher resolution (ie. can act as a ground truth for the spots' composition)?

In your paper, there is the 'Spatial registration of snRNA-seq' section, so I am wondering if the single cell/nucleus spatial information is available for download?

Sorry if this is explained somewhere... I am keen on using your dataset if the (ground truth) single cell spatial information is available (for paired Visium spot data) -- unfortunately, given my limited computer capacity, downloading many datasets to explore is not possible, therefore I was hoping to better understand the datasets before downloading.

Thanks!

spatialLIBD • 110 views
1
Entering edit mode
Last seen 15 days ago
United States

Hi,

The SingleCellExperiment object you can download with fetch_data(type = "sce") pre-dates the creation of SpatialExperiment. Nowadays, you should really only use fetch_data(type = "spe"). The SCE object was our hacked solution prior to SPE. spatialLIBD automatically re-casts it from SCE to SPE when you use fetch_data(type = "spe") thanks to sce_to_spe(). https://bioconductor.org/packages/release/data/experiment/vignettes/spatialLIBD/inst/doc/spatialLIBD.html#re-shaping-your-data-to-our-structure goes into more detail into the original SCE structure, but again, use the SPE one.

Note that the SpatialExperiment pre-print is different from the final published version, since reviewers asked to change a few internal things, so you should read http://dx.doi.org/10.1093/bioinformatics/btac299.

Best, Leo