search

daviddaiweizhang / istar

GNU General Public License v3.0
74 stars 6 forks source link

Question about istar application on 10× Visium data #5

Open watermelontreesjs opened 5 months ago

watermelontreesjs commented 5 months ago

istar is a really nice work!

But I have the following concerning about applying istar to 10× Visium data. ① Usually, the output of 10× Visium data include a folder named spatial, which containing issue_hires_image.png image with pixel size 2000×2000 (6.5 mm Capture Area). And according to (https://www.10xgenomics.com/support/software/space-ranger/latest/analysis/outputs/spatial-outputs), Each capture area has 78 spots per line. Without considering the spot-to-spot interval distance, the pixel size of each spot is a maximum of 25 and even smaller because each spot is not tightly close to each other.

② I noticed that istar partitioned the image into 16×16 pixel. This 16×16 pixel is so similar to the pixel distance of the spot that I am concerned that istar predictions may not give the users much information about the gene prediction on the slice. The clustering of each spot using only raw gene expression sequencing may be very similar to that provided by istar.

I would like to consult the authors on the above question and ask everyone who see this issue and if there is a better way to apply istar to 10× Visium data.

LenisLin commented 2 months ago

@daviddaiweizhang

yaosicong1999 commented 3 weeks ago

@watermelontreesjs @LenisLin Because the _issue_hiresimage.png provided in the spatial folder is only in high resolution, not full resolution. The full-resolution image is often in .tiff format that you may (or may not) find with other files. It should appear under Input Files for datasets publicly available on https://www.10xgenomics.com/datasets if the full-resolution image is available.