boomstarcuc / genesegnet Goto Github PK

Congratulations on the publication!

I am interested in trying it out with my own data and I have some questions:

1. Is the image size 256 x 256 the one you recommend?
2. Can one also use cell border staining or is nuclei staining prefered?
3. How exactly does my input need to be formated? Just Spot X,Y, ground truth and staining images?
4. Any recommendation on generating the ground truth masks? Did you just use another segmentation algorithm for this?

Thank you!

Hi,
Congratulations for this amazing work.
I ran the NSCLC model on nanostring NSCLC data. The output seems to have visible lines in the borders of overlapping patches. Do you have any suggestion? I tried the NSCLC model on hippocampus data and it did not have such a problem. Do you have any suggestion?

Dear authors,

Thanks for your effort! GeneSegNet truly impresses.

However, when I attempted to utilize it with my personal data using the preset models, I encountered some small issues. In particular, in "slidingwindows_gradient.py," the code expects "image," "label," and "spot," while our folder names were created as "images," "labels," and "spots," causing issues at lines 248, 249, 250. Additionally, in line 304, I propose modifying the code to "spot_list.append([int(float(splits[0])), int(float(splits[1]))])" instead of "spot_list.append([int(float(splits[1])), int(float(splits[2]))])", as they should be x, y coordinates. Could you please validate these problems by testing on various datasets for resolution?

Moreover, the current instructions seem a bit simplistic. After exploring GeneSegNet for several hours, I'm still uncertain about the specific output files it generates. Could you provide more comprehensive details regarding the output structure? Additionally, if possible, providing a vignette would be immensely helpful.

Thanks,
Yiqian

Hi team, nice work. Do you have script to reproduce the cell boundary figure showing in your paper?

I'd appreciate a walkthrough that includes the steps for reading the files in the output directory, processing the data within, and then using that data to draw the cell boundaries.
Thanks!

Dear Authors

Thank you for creating this promising tool for cell count segmentation in spatial transcriptomics. Though I do face some issues to prepare the input files with the limited given information on preparation...
I am currently engaging with GeneSegNet for analyzing spatial transcriptomics data and require assistance in preparing inputs from Visium AnnData and Ilastik probability maps (as Validation).

Context and Issue:

• Objective: To utilize GeneSegNet for analyzing Visium platform data, integrating it with cell segmentation probability maps generated through Ilastik.
• Challenge: I am encountering difficulties in correctly formatting and aligning Visium AnnData with Ilastik probability maps for cell nuclei segmentation.

Specific Questions:

1. What is the recommended approach for preparing Visium AnnData for compatibility with GeneSegNet?
2. How should I transform Ilastik probability maps to align them correctly with Visium data?
3. Are there specific preprocessing steps or format conversions required for the probability maps before feeding them into GeneSegNet?
4. How should the image be chunked and transformed for the model?

Any guidance, documentation references, or examples of processing such data for GeneSegNet would be immensely helpful.

Thank you for your time and assistance! :)