Skip to content

The repo of the "Multi-omics alleviates the limitations of panel-sequencing for cancer drug response prediction" manuscript

Notifications You must be signed in to change notification settings

BIMSBbioinfo/multiomics_vs_panelseq

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Multiomics vs Panelseq

Supporting repository for the manuscript of the same name. Here, one can find scripts and instructions to recreate the work partly or as a whole.

Downloading the data

get_data.sh script is here to access the data used in the work. Example usage:

git clone https://github.com/BIMSBbioinfo/multiomics_vs_panelseq.git
cd multiomics_vs_panelseq
bash get_data.sh -o <out_path>

Building drug response prediction models

After the selected data were downloaded, one has multiple paths to follow. To generate models from the processed data the code section below would be sufficient:

usage:
Rscript src/model_DR.R <dataset> <path_to_data_dir> <path_to_write_models> <number of runs> <number of cores>
example (building models using CCLE or beatAML datasets, running 1 modeling run per drug using 10 cores:
bash get_data.sh -o ./
Rscript src/model_DR.R CCLE ./data ./results 1 10

A modeling run for each drug can be repeated multiple times (we do this for PDX samples):

Rscript src/model_DR.R PDX ./data ./results 20 10

Once the models are built, performance metrics and variable importance metrics can be extracted and saved with the following

Rscript ./src/analyse_models.R ./results 20 #using 10 cores to process results

Producing manuscript figures and tables

Finally, figures and tables can be collated:

Rscript ./src/collate_figures_tables.R ./results <get_data.sh out_path>

About

The repo of the "Multi-omics alleviates the limitations of panel-sequencing for cancer drug response prediction" manuscript

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published