Setting up Qfeatures - QUBO_Feature_Selection

0. Pre-Installation Steps (Python)

To set up the environment for running the QUBO notebook, follow these steps:

0.1. Create and Activate the Conda Environment

conda env create -f environment.yml
conda activate dwave

0.2. Install the qfeatures Package

pip install -e .

0.3. Install Jupyter Kernel (Linux)

python -m ipykernel install --user --name dwave

0.4. Set up D-Wave API

1. Visit D-Wave Leap and find your Solver API Token.
2. In your terminal, configure your D-Wave API client:

   dwave config create

3. When prompted, paste your Solver API Token and hit Enter.

0.5. Add MATLAB Path

Add the following directory to your MATLAB path:

C:\Users\ssromerogon\Documents\vscode_working_dir\QUBO_Feature_Selection\qfeatures-src-v0.2_matlab

1. Options for Computing QUBO Matrix

You have three options for generating the QUBO matrix:

1.1. Option 1: Generate QUBO Matrix from MATLAB

• Run the script qfeatures_driver.m located in ../../../qubo_fs_matlab/efficient_differentiation/.
• Copy the generated files genes.csv and qubo_matrix.csv to the current local directory.

1.2. Option 2: Use a Pre-Computed QUBO Matrix

• Download the pre-computed matrix from the following link: Google Drive - Pre-Computed QUBO Matrix

1.3. Option 3: Run the Stand-Alone Python Notebook

• After installing the qfeatures package (pip install -e .), run the stand-alone notebook construct_qubo_python.ipynb in this directory.
• Download the prepared single-cell data from the following link: Google Drive - Single-Cell Data File: Data_hESC_EC_day1_5000g_filtered_feature_bc_matrix_h5.h5ad

2. Running the QUBO Notebook

• Open qubo_solver_from_MATLAB_Q.ipynb in Jupyter Notebook.
• Run all cells in the notebook.
• The selected features will be saved in filt_df_QA.csv.

2.1 Running in matlab

• Open qfeatures_driver.m located in ./qubo_fs_matlab/efficient_differentiation/ and execute it (NOTE: install Quantum Computing matlab package)
• Cross validation available in qfeatures_driver_cross_validation.m in ./qubo_fs_matlab/efficient_differentiation/