-
Notifications
You must be signed in to change notification settings - Fork 1
Setup
Install Miniconda with Python 3.7. Download this repo content into some directory (referred to as REPO_PATH below), and create a new Conda environment with this command:
$ conda env create -n YOUR_CONDA_ENV_NAME -f [REPO_PATH]/env/environment.ymlBefore running the code below, make sure to activate the environment and modify your PYTHONPATH variable:
$ conda activate YOUR_CONDA_ENV_NAME
$ export PYTHONPATH="[REPO_PATH]/src:$PYTHONPATH"
Open the file [REPO_PATH]/src/settings.py and modify the BASE_DIR to point to a directory in your computer (it will be created if doesn't exist).
To obtain the raw/base results (output files from S-PrediXcan, S-MultiXcan, TORUS and fastENLOC), you can run all again from scratch, or download the results we generated.
If you want to run everything from scratch, check out these directories:
If you just want to run the postprocessing scripts on these results, you can avoid running everything from scratch by downloading raw results. To automatically download all results and needed data for postprocessing, run these commands:
$ cd [REPO_PATH]/scripts/100_postprocessing/
$ python 00_setup.py all --extract
...
smultixcan_100001_raw_ccn30.tsv.gz: downloading... hash ok... done
smultixcan_100002_raw_ccn30.tsv.gz: downloading... hash ok... done
...
spredixcan-100001_raw.tar.bz2: downloading... hash ok... uncompressing ...done
spredixcan-100002_raw.tar.bz2: downloading... hash ok... uncompressing ...done
...For S-PrediXcan, fastENLOC and TORUS results, archives (.tar.bz2) need to be unpacked. The command above will do that automatically since the --extract parameter was used. This will only extract the archive but not delete it, what needs more disk space. If you want to delete the archive after unpacking, use the --extract-and-delete instead.