master-thesis#
This repository contains the code for my master thesis.
Data#
Download the original dataset here.
Set up environment#
Install Miniconda. conda env create -f environment.yml
. Or install dependencies as below.
conda install -c conda-forge notebook
conda install -c stellargraph stellargraph
pip install pydot pydotplus graphviz
pip install scanpy[louvain,leiden]
pip install pynndescent
pip install phate
conda install -c conda-forge ipywidgets
conda install pytorch==1.5.1 torchvision==0.6.1 cpuonly -c pytorch
pip install torch-scatter==2.0.5 -f https://pytorch-geometric.com/whl/torch-1.5.0+cpu.html
pip install torch-sparse==0.6.6 -f https://pytorch-geometric.com/whl/torch-1.5.0+cpu.html
pip install torch-cluster==1.5.6 -f https://pytorch-geometric.com/whl/torch-1.5.0+cpu.html
pip install torch-spline-conv==1.2.0 -f https://pytorch-geometric.com/whl/torch-1.5.0+cpu.html
pip install torch-geometric==1.6.0
Some useful commands:
conda activate
jupyter notebook --ip=127.0.0.1 --no-browser
./vips-dev-8.10/bin/vips.exe dzsave ./results/nuclei_heart.tif --tile-size=254 --overlap=1 --depth onepixel --suffix .jpg[Q=90] heart
Scripts#
All the scripts are self-contained and can run in notebook mode. The function of each script and the corresponding figures produced in the thesis are listed below. Refer to the scripts for more descriptions.
preprocess.py
: fig 7, s5graphsage.py
,dgi.py
,gcn.py
: StellarGraph (not used in the thesis)pyg-sage-rw.py
,pyg-sage-dgi.py
,pyg-gcn-rw.py
,pyg-gcn-dgi.py
: PyGcluster.py
: Louvain, PAGA, UMAPplot-umap.py
: fig 2, 3hm.ipynb
: fig s1, 4, 8, s2paga.py
: fig 5, s3analysis.ipynb
: fig 6, 9, 10, 11pyg-ge.py
: fig s4stitch.py
: for TissUUmapsrun.sh
: simple logging