OV_Histotypes

Classification of Ovarian Cancer Histotypes

Generate RSF

The Report of Statistical Findings (RSF) can be generated by following the steps below:

1. Clone the repo to your local machine and open RStudio
2. Two local packages otta and ottaOvca need to be installed locally. They are located on the I drive (\\files.ubc.ca\team\gpec\i) at I:/data repo. Copy the packages and put them at the path RENV_PATHS_LOCAL (found in .Renviron). Install using renv::install() and use as input the file path to the source package.
3. For the rest of the package dependencies, run renv::restore() to restore all packages used in this project into a private library on your local machine.
4. Restart the RStudio session and the Build tab should now have a Build Book button. Click that button to generate GitBook (HTML) and PDF formats of the RSF, or use the dropdown menu to generate a specific format only.
5. The updated report files are generated in docs/. Push these files to the GitHub remote. Note that the files need to be pushed to the master branch to update the report.
6. RSF is generated at URL https://talhouklab.github.io/OV_Histotypes/

Classifier Pipeline

The classifier pipeline can be run by following the steps below:

1. Clone the repo to a GSC server. One of the gphosts is a stable server to use as it runs the relatively newer CentOS 7 operating system (gphost01, ..., gphost04).
2. Follow Steps 2 and 3 of [Generate RSF] to install R package dependencies. You'll need to use sftp/scp/rsync to transfer the local packages to the server.
3. SSH to the numbers server to submit SLURM jobs. This is the job management system that allows us to fire away a large number of scripts to worker nodes.
4. Open a new tmux session to run the pipeline in so that the SSH session doesn't timeout if it is running for a long time.
5. Navigate to pipeline/ and run make all
6. The outputs will be in output directory specified in pipeline/assets/params.sh. You can also modify parameters in this script as necessary to change the pipeline.