Tree of Mutually Exclusive Oncogenes
The notebook demo.ipynb provides a comprehenisve illustration of various structural moves using an example.
The experiments can be reproduced using syn_datagen.py and syn_postproc.ipynb
The data can be downloaded from GDAC Firehose. We use Mutation_Packager_Calls (MD5) files as the input in our preprocessing steps. The IntOGen lists of potentially driver genes can be downloaded from intogen website. For Glioblastoma multiforme for instance, the table including the list of driver genes can be downloaded from here.
Use gdac_preproc.py to prepare the input csv file for the algorithm.
$ python tomexo.py -i input.csv -o output --n_chains 10 --n_mixing 5 --n_samples 100 --pp 5
- n_chains default is 10
- n_mixing default is 0
- n_samples default is 100000
- pp is the prior's power variable, with default value of 5
The script in gdac_postproc.py performs a basic postprocessing and generates a pdf file of the final progression model.