driver_neighbours

Code for the analysis of neighbours of cancer drivers.

Raw data files used in this analysis are available on Zenodo. Extract data.zip to the root of the code directory.

To reproduce the figures in the manuscript use bct_wct_figures.R. All the output files needed to reproduce the figures are already available in the Zenodo repository.

To reproduce the entire workflow, first run the python notebooks 1 to 4 (as indicated below) and then the R code.

Python code

To reproduce python code, use the environment.yml file to create a virtual environment using conda, pixi or other similar tool.

The work is divided into 5 jupyter notebooks:

1. get_PPIs.ipynb: this is the first part of our work. It retrieves and processes protein-protein interactions (PPI) data.
2. preprocessing.ipynb: This notebook uses the processed PPI data and raw TCGA data to create files for BCT and WCT analysis.
3. pathway_analysis.ipynb: Computation of shortest path lenghts between drivers and neighbour genes.
4. bct_analysis.ipynb: Runs BCT analyses. It also includes the tumour mutational burden calculation.
5. opentargets_analysis.ipynb: Contains the last part of our work. Run it after the R code. Using the Open Targets Platform, it searches for potential therapeutic targets among candidate neighbours.

R code

R packages needed:

1. doParallel
2. doSnow
3. ggplot2
4. ggpubr
5. arrow
6. tictoc
7. Coselens
8. disgenet2r
   

To reproduce the figures in the manuscript:
Run bct_wct_figures.R

To run the tumour-normal paired sample analysis and compare it with the signalling pathway analysis:
Run paired_analysis.R
Note: code is parallelized. Change nclust accordingly to the number of available cores.

To run wct analysis (including analysis of shuffled datasets:
Run wct_computation.R
Note: code is parallelized. Change nclust accordingly to the number of available cores. Tests with permuted data take long time to run.

To analyse bct and wct results by neighbour and by driver:
Run bct_wct_analysis.R

To run the validation of wct interactions using Coselens:
Run coselens_validation.R

bct_wct_functions.R contains accessory functions used in the remaining R scripts.