cell_labelling.md

Cell Based Labelling Algorithms

• ./cell_based_program/CV.R
  • Generates the 5-fold cross validation groups as .Rdata files
  • Inputs: A file containing labels for each cell
  • Outputs: An Rdata file containing cross validation groups
  • Arguments
    • Path to the genes x cells counts file
    • Path to the folder in which to save the Rdata file
    • Path to the folder containing the cell labels
• ./cell_based_program/CV_r2py.py
  • Converts the Rdata files from CV.R to pkl files for the python methods
  • Inputs: The Rdata file
  • Outputs: The Pkl file, in the same directory as the Rdata file
  • Arguments: Path to the folder containing the Rdata files
• ./cell_based_program/R_methods.R
  • Runs all R based methods on a single dataset
  • Inputs: A counts matrix, labels file, and Rdata CV file for the dataset
  • Outputs: Predictions in ./cell_based_program/output/<dataset>
  • Arguments:
    • The name of the dataset
    • The path to the folder containing the counts matrix, labels, and Rdata
• ./cell_based_program/Python_methods.py
  • Runs all python based methods on a single dataset
    • Lambda and scVI have their own scripts
  • Inputs: A counts matrix, labels file, and pkl CV file for the dataset
  • Outputs: Predictions in ./cell_based_program/output/<dataset>
  • Arguments:
    • The path to the folder containing the counts, labels, and pkl
    • The name of the dataset
  • DEPENDENCIES IN ./cell_based_program/OtherMethods_packages.txt
• ./cell_based_program/run_LAmbDA.py
  • Runs LAmbDA on a single dataset
  • Inputs: A counts matrix, labels file, and pkl CV file for the dataset
  • Outputs: Predictions in ./cell_based_program/output/<dataset>
  • Arguments:
    • The path to the folder containing the counts, labels, and pkl
    • The name of the dataset
• ./cell_based_program/run_scVItool.py
  • Runs scVI on a single dataset
  • Inputs: A counts matrix, labels file, and pkl CV file for the dataset
  • Outputs: Predictions in ./cell_based_program/output/<dataset>
  • Arguments:
    • The path to the folder containing the counts, labels, and pkl
    • The name of the dataset
  • DEPENDENCIES IN ./cell_based_program/scANVI_packages.txt
    • This requires a different version of tensorflow, it is recommended to use a virtual environment to manage it.
• ./cell_based_program/other_scripts/results_table.ipynb
  • Gathers the predictions from all methods into a single file
• ./cell_based_program/other_scripts/time_sim.py
  • Gathers the running times into a single file
• ./cell_based_program/other_scripts/time_bar_plot.ipynb
  • Makes bar plots of the running time as seen in our supplemental figures
• ./cell_based_program/other_scripts/time_coefficient_plot.ipynb
  • Makes the dataframe needed for the time coefficient plot generated in main_figures.ipynb
• ./cell_based_program/other_scripts/heatmap.ipynb
  • Makes the dataframe needed for the running time heatmap generated in main_figures.ipynb