This repository contains the code to reproduce the plots used in the SPAGHETTI paper. The main SPAGHETTI program can be found at: https://github.com/schwartzlab-methods/spaghetti
Please first download the datasets used to train and validate the model:
- LIVECell (used for both training and validation): http://livecell-dataset.s3.eu-central-1.amazonaws.com/LIVECell_dataset_2021/images.zip
- PanNuke (used for training only): https://warwick.ac.uk/fac/cross_fac/tia/data/pannuke
- C2C12 Culture Media (used for validation only): https://osf.io/ysaq2/
- DIC ALFI dataset (used for validation only): https://springernature.figshare.com/articles/dataset/ALFI_dataset_final_/23798451?backTo=/collections/ALFI_Cell_cycle_phenotype_annotations_of_label-free_time-lapse_imaging_data_from_cultured_human_cells/6436958
- Brightfield dataset (used for validation only) from
data/: https://zenodo.org/records/8415315
Please ensure that all images are organized by the cell type or culture media. For example, your LIVECell folder should look something like this after your pre-processing:
livecell/
- A172/
- all_images_of_A172_cells
- BT474/
- all_images_of_BT474_cells
...
In addition, please also download the cell viability dataset, which can be used as is. You can download it from https://figshare.com/articles/dataset/VirtualStaining_Dataset/21971558/1?file=38982755
All scripts are located at script_for_figures/. Before running the scripts, since there will be a lot of outputs and inputs being generated, make sure you go to each bash script file and change the path to the corresponding inputs and outputs based on the comments in the bash script. Please also read the argument for each .py files to be run to understand what argument is required.