GBM Hackathon: Data Loading

Data loading API and data visualization of GBM Hackathon data.

Table of Contents

• Installation
• Starters
• Credit

Installation

1. Start by initialising Conda.

conda init bash && bash

2. Change the default conda environment location. This ensures it is not necessary to reinstall the environment each time you restart a Notebook instance. This command needs to be rerun each time you restart the instance.

conda config --append envs_dirs /home/ec2-user/SageMaker/.conda/envs

3. Create the environment.

conda create -n gbmhackathon -y python=3.10
conda activate gbmhackathon

4. Optional, fork the repository If you aim to share your code with your team members, consider doing a fork or create a copy of the actual repository with a new name.

5. Clone the repository and install the dependencies.

cd /home/ec2-user/SageMaker
git clone https://github.com/owkin/gbm_hackathon.git
cd gbm_hackathon
make install-poetry
make install-all
pre-commit install

Note that this also makes your environment visible in Jupyter.

Starter Notebooks

To help you start working quickly on the data, we provided you with the following starter Notebooks.

• notebooks/data_loader_demo.ipynb Will help you understand the data loading and which data sources (modalities) are available.
• notebooks/visium_starter_mosaic.ipynb Will help you load and explore the Visium data. This includes performing data normalization, unsupervised clustering, and quantifying cell population and transcription factor activities.

We hope these Notebooks will help you with your data analyses!

Credit

We would like to thank the following contributors to this project: Lucas Fidon, Quentin Bayard, Alex Cornish and Valérie Ducret.