Project for special course at DTU Compute together with Maximillian Fornitz Vording in collaboration with Pers Lab.
Using deep learning methods on single-cell sequencing data, we want cluster cells and potentially classify their cell types.
The model has been implemented in Python using the Theano, Lasagne, and Parmesan modules. In addition, NumPy and matplotlib are also used to for computations and making figures.
Data are expected to be in subdirectory called data. All other necessary subdirectories are created as needed.
For how to run the model, run ./src/main.py -h.
The shell script run.sh has been supplied to make it easier to specify the arguments to ./src/main.py.
Most of the model specification has been taken from implementation of the variational auto-encoder from the Deep Learning course at DTU.