-
mainCode.ipynb - containing the main code of our model (need to change the input file path)
-
1M_neurons_neuron20k.h5 - the input dataset we used (can be download from https://support.10xgenomics.com/single-cell-gene-expression/datasets/1.3.0/1M_neurons)
-
Final_Paper - final project report
-
Dataset - containing the input dataset we tried
-
Simulate.ipynb - code for generating simulated dataset,
nGenes
: gene numberbatchCells
: cell numbergroup.prob
: proportion of each cell population
-
GiniClust3.ipynb - the code file for implementing GiniClust3 (used for the comparison with our model)
-
desc.py, network.py, SAE.py (desc code download from https://github.com/eleozzr/desc, Make sure they are in the same folder as the mainCode.py file)
-
Simulated_dataset - folder containing all the simulated dataset (with different proportion of rare cell)- This folder is too big thus cannot be uploaded, but the simulated dataset inside can be generated by running Simulate.ipynb
-
Result - containing output images and scores.
- giniclust3 (https://github.com/rdong08/GiniClust3)
- Installation:
pip install giniclust3
- Installation:
- desc (https://github.com/eleozzr/desc) (no need to install this package, just import SAE.py and network.py)
- Installation:
pip install desc
- Installation:
- leidenalg (https://github.com/vtraag/leidenalg)
- Installation:
pip install leidenalg
- Installation:
- Tensorflow 2*
Import the dataset you want to run with, and then run the code.
Make sure the mainCode.ipynb, network.py and SAE.py are all in one folder
comment or delete the the code below if you run the code locally
from google.colab import drive
drive.mount('/content/drive')