https://github.com/jkobject/hackathon_bio_ai
https://github.com/jkobject/hackathon_bio_ai
Last synced: 5 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/jkobject/hackathon_bio_ai
- Owner: jkobject
- License: mit
- Created: 2024-12-06T08:22:18.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-12-23T19:30:11.000Z (over 1 year ago)
- Last Synced: 2025-12-13T03:56:21.986Z (7 months ago)
- Language: Jupyter Notebook
- Size: 54.1 MB
- Stars: 2
- Watchers: 1
- Forks: 0
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Bio AI Hackathon 2024
Functionality:
- Getting embeddings using CONCH 1.5 for patches and TITAN for slides
- Scripts to run [scPRINT](https://github.com/mahmoodlab/HEST) on spatial xenium datasets to get cell types and full gene expression
- Downloader / converter of HEST1K to spatial data files in Lamindb
- DataLoader for spatialdata
- Extract embeddings from TITAN and store in .obsm() and ask questions
- Buttons and Textboxs to Napari to ask questions and search for patches
### 1. Install the package in editable mode
```bash
git clone git@github.com:jkobject/hackathon_bio_ai.git
cd hackathon_bio_ai
conda create -n st_challenge_dev python=3.10
conda activate st_challenge_dev
pip install -e .
```
### Install the package
```bash
conda create -n st_challenge_prod python=3.10
conda activate st_challenge_prod
pip install git+https://github.com/jkobject/hackathon_bio_ai.git
```
### 4. View an image
```bash
conda activate theremia
python -m st_challenge.scripts.viz_slice INT1
```