search

seqscope / ficture

https://seqscope.github.io/ficture/
Other
23 stars 2 forks source link

transcript-level output #7

Open ruqianl opened 4 months ago

ruqianl commented 4 months ago

Great work! I really like the simulation module too.

I am wondering with FICTURE, is there any way to output the transcript-level results instead of pixel-level?

Transcript-level results for each transcript in the input transcripts.tsv

Thanks, RL

hyunminkang commented 3 months ago

Hi @ruqianl -- sorry for the delayed response. I believe that @Yichen-Si can give a better answer to this, but this is what I can answer:

With the current version, the pixel-level output (.pixel.sorted.tsv.gz) should have a resolution high enough to associate it with individual transcripts. So you can "join" the transcript-level output with the pixel-level output using simple nearest neighbor search.

We have a plan to provide such a script to "join" the two datasets in the near future, but it has not been implemented into the package yet. We will update in this thread when this is done.

hyunminkang commented 1 month ago

Hi @ruqianl -- this is a bit of late follow-up, but we have published a tool to combine the transcript-level output and pixel-level output so that it can be effectively "joined" together.

Please see the example usage in the spatula documentation at the following URL:

https://seqscope.github.io/spatula/tools/join_pixel_tsv/

We will appreciate if you could let us know whether it worked for your own dataset or not.