aertslab / spatial_fly_website

0 stars 0 forks source link

Cell types #1

Open sthawinke opened 10 months ago

sthawinke commented 10 months ago

Hi, thanks for an exemplary data availability!

I would like to use your dataset for an analysis, and have found the molecule coordinates allright, but I was wondering where I can find the cell types as mentioned in e.g. section "Building a spatial tissue atlas of the fly body" in the manuscript?

Many thanks!

jjans5 commented 10 months ago

Hi sthawinke,

Would you be interested in getting the cell segmentation or the annotation of mRNA spots to cell types (either directly or through the grid assignment)? We are currently already working on a scope session that will contain the cell segmentation results (https://scope.aertslab.org/#/Spatial_Fly/ ). I will look to upload the annotation of mRNA spots as separate datafiles to the website.

Cheers!

sthawinke commented 10 months ago

Hi jjans,

I am looking for both, although in the first place for the annotation of mRNA spots to cell types indeed. In the meanwhile I will have a look at the loom files too.

Thanks,

Stijn

sthawinke commented 6 months ago

Hello jjans,

I was trying to read in the loom files, but loompy complains that

... does not appear to be a valid Loom file according to Loom spec version '3.0.0'

Which version of loom are you using? Or could you upload the loom files after the backwards-incompatible changes of v 3.0.0?

Many thanks,

Stijn

KrisDavie commented 6 months ago

Hey Stijn,

The loom files formatted for SCope follow our own format that was established before the checks were put in place in loompy >=3.0.0. However, you can still load these files with loompy >=3.0.0, you just have to pass the validate=False argument to the loompy.connect call.

Cheers,

Kris

sthawinke commented 5 months ago

Hi Kris,

Thanks for sorting that out. I'm still not sure how to map the molecule coordinates onto cell segmentation or cell type annotation, is this info present in the loom files? There are some cell type and x-y coordinates in the loom file, but the coordinates have negative values too and seem to be from some embedding.

Thanks a lot,

Stijn