Cancer Mutation Finder CLI

This Python 3 CLI app retrieves data from the cancer bio portal and does a rudimentary analysis of mutation frequency of the supplied gene symbols. Gene symbols are limited to the 1000 common genes listed in app/data/gene_id_symbol_map.csv. Currently the study ID is firm-coded to "gbm_tcga_gistic".

Getting Started

Dependencies:

• Internet access
• Python 3
• pip

Installation

cd DFCI
virtualenv venv  # Recommended
source venv/bin/activate # Recommended
pip install -r requirements.txt

Note the shebang line in app/main.py may not work for all systems. In that case run main.py as a script. I.E. $ python3.6 main.py TP53.

Using Command Line Tool

Each parameter is a gene name i.e. TP53 or BRCA1.

$ ./main.py TP53
TP53 is mutated in 22.34% of all cases.
TP53 is copy number altered in 1.83% of all cases.

The user can also be able to execute the program with as many genes as desired genes. For example:

$ ./main.py TP53 MDM2 MDM4 MXA XPA
TP53 is mutated in 22.34% of all cases.
TP53 is copy number altered in 1.83% of all cases.
MDM2 is mutated in 23.81% of all cases.
MDM2 is copy number altered in 9.16% of all cases.
MDM4 is mutated in 30.40% of all cases.
MDM4 is copy number altered in 9.52% of all cases.
XPA is mutated in 26.74% of all cases.
XPA is copy number altered in 0.37% of all cases.
The gene set is altered in 61.54% of all cases.

And finally, because life is short, the --toJSON flag results in JSON being sent to standard out. Note I used jq to format the output for legibility in this example.

$ ./main.py TP53 MDM2 MDM4 XPA --toJSON | jq
{
  "geneset": {
    "mutated": 168,
    "totalCount": 273
  },
  "TP53": {
    "NA": 0,
    "singleCopy": 56,
    "noCopy": 5,
    "multiCopy": 0,
    "noChange": 212
  },
  "MDM2": {
    "NA": 0,
    "singleCopy": 40,
    "noCopy": 0,
    "multiCopy": 25,
    "noChange": 208
  },
  "MDM4": {
    "NA": 0,
    "singleCopy": 57,
    "noCopy": 0,
    "multiCopy": 26,
    "noChange": 190
  },
  "XPA": {
    "NA": 0,
    "singleCopy": 72,
    "noCopy": 0,
    "multiCopy": 1,
    "noChange": 200
  }
}

Known Issues

• There are (almost) no test. Unit test and at least one integration test would be awesome and wouldn't take very long.
• Executing the script from outside of /app resulting in incorrect path to /data. This could be fixed pretty easily with:

  import os
  dirname = os.path.dirname(__file__)
  filename = os.path.join(dirname, 'relative/path/to/file/you/want')
  

• Single genes should also return the total % of mutations. I.E. "Total % of cases where TP53 is altered by either mutation or copy number alteration: 30% of all cases." for ./main.py TP53
• The error messages are uninformative and should tell a user why the app is failing. Usually this is because the gene symbol isn't recognized.
• None of the functions have doc string. Might not be that big of a deal since they're not designed for export but it would help with maintenance.

Future Features

• Specifying a study ID could add a lot of value and is trivial.
• Dockerizing the whole app while definitely overkill would solve shebang line issues and negate the need to install requirements.txt.
• If a gene symbol isn't found in app/data/gene_id_symbol_map.csv the app should look for it online.
• Gene results should be cached. (I.E. app/data/<study_id>/<gene_id>)