Yu-AnChen/quick-look

Preview whole-slide scans using ngff and napari

Quickly tile whole-slide scans into a pyramidal images for holistic image review and data-driven decision-making in highly-multiplexed tissue imaging.

NOTE: Whole-slide scans from RareCyte imagers are supported. Processing file formats from other vendors are feasible but is not supported at the moment.

Installation

Installing palom in a fresh conda environment is recommended. Instruction for installing miniconda.

Install quick-look on the computer that does the tiling

In most cases, this is also the computer that acquires the images.

conda create -n quick-look -c conda-forge python=3.10 numpy scipy matplotlib networkx scikit-image=0.19 scikit-learn tifffile zarr pyjnius blessed tqdm fire watchdog joblib pywin32 git

conda activate quick-look

python -m pip install git+https://github.com/Yu-AnChen/quick-look.git

Install napari and napari-ome-zarr to view the pyramidal images

conda create -n napari -c conda-forge python=3.10 pyqt napari

conda activate napari

python -m pip install napari-ome-zarr

Using quick-look

Usecase 1: process scans "by hand"

quicklook process -i <path/to/input/directory> -o <path/to/output/directory>

quicklook process processes folders or files in the input directory (<path/to/input/directory>). The folders must contain one .rcpnl file and optionally one .rcjob file; thie files must be a .rcpnl files. On Windows computers, the folders and files can be "shortcuts". The user doesn't need to move the actual files to the input folder.

Example input, containing one .rcpnl file and one folder:

C:\quick-look\input
│   LSP001@20230922_142033_633584.rcpnl
│
└───LSP002@20230922_200722_567967
        LSP002@20230922_200722_567967.rcjob
        LSP002@20230922_200722_567967.rcpnl

Command executed

quicklook process -i C:\quick-look\input -o C:\quick-look\output

Example output, two .ome.zarr are generated:

C:\quick-look\output
├───LSP001@20230922_142033_633584.ome.zarr
└───LSP002@20230922_200722_567967.ome.zarr

Usecase 2: monitor a directory and tile a scan right after its file is created

quicklook monitor -i <path/to/onput/directory> -o <path/to/output/directory>

Use quicklook monitor command to monitor a folder (<path/to/onput/directory>) on a computer. Everytime an user paste shortcuts (of folders or files as described in usecase 1) into the folder, tiling will be launched, the results will be written to <path/to/output/directory>.

Review the tiled whole-slide scans

quick-look writes out NGFF v0.4 files. Using Napari with napari-ome-zarr is currently recommended for opening quick-look outputs. Other tools that might be able to open the images are listed here.