wehos / CellT

An offical repo for paper "Single-cells are Spatial Tokens: Transformers for Spatially Resolved Transcriptomics Data Imputation"
16 stars 2 forks source link

Tutorial #2

Open jayypaul opened 11 months ago

jayypaul commented 11 months ago

Hi!

I enjoyed the paper and am really interested in trying out the method. Would you be willing to write a short tutorial, or give some advice on which functions to use to perform imputation with the recommended default settings in the paper? I have multiple Cosmx datasets as well and I'd like to evaluate the performance of the method.

Thanks!

Jordan

wehos commented 11 months ago

Hi Jordan!

Thank you very much for your interest. I am more than happy to provide a tutorial. We are also planning a revision of the paper. I'll try my best to get it done before Jan 2024.

Best, Hongzhi

jayypaul commented 11 months ago

That would be awesome! Any recommendation for which functions should be used and in what order would be greatly appreciated.

Happy holidays!

jayypaul commented 10 months ago

Hi again,

Just wanted to follow up and see if you could provide some advice on which functions should be used to run your method? I have Cosmax data collected from prostate tissue, from multiple fov's. I'm mainly interested in testing it on one fov first to see how it performs. The fov's are of different parts of the prostate affected by cancer. Some are in tumor regions, others are stroma, others benign epithelial, etc. so I'm expecting to only be able to use fov's collected from tumor regions as training data to perform imputation on those tumor regions for example. For a given patient, there's only 3 fov's collected from tumor, 3 from benign, and other tissues even less.

Any advice on how to properly use your framework to perform imputation given the data set up, that would be very helpful!

Happy new year.

Jordan

wehos commented 10 months ago

Hi, Jordan!

Sorry for the late response. I was busy releasing a tutorial for our pre-trained model CellPLM (which is a variant version of this SpaFormer model). Now I finally get it done and can come back to SpaFormer development. While we are working on SpaFormer now, could you please take a look at CellPLM tutorial to see if you are also interested in it?

Regarding your requirement, if you are comfortable about that, could you please send me an email at wenhongz@msu.edu so that we can have a more detailed discussion?

Thanks again for your interest!

jayypaul commented 8 months ago

Hi @wehos ! Hope all is well!

Just wanted to follow up on this thread. I'm sure you're busy, but just wanted to get a sense of whether there's a planned timing for release of a tutorial coming up?

Best,

Jordan

ATar260 commented 1 month ago

Hi is there any updates on this ?