Color Genomics risk models.
The Claus risk model is a model of breast cancer risk from familial data. The paper publishes tables that map risk percentages based on each family member's relationship and cancer onset ages.
This project is a Python implementation of this paper. Given a patient's family history of cancer diagnosis, it returns the lifetime risk of the patient developing breast cancer assuming he/she is currently cancer-free.
Given the cancer onset ages of each family relationship, the function looks at the risk scores derived from all applicable Claus tables and selects the highest risk score.
Example: For a patient whose mother and two aunts have developed cancer. The function looks at both the mother + aunt table as well the two second degree relatives table and does each calculation. Depending on the ages of diagnosis, either table can result in higher lifetime cancer risk score.
A key part of calculating remaining risk in the equation specified by the paper is calculating the risk of developing cancer up until the patient's current age and conditionally removing that probability from lifetime risk (as we know the patient is currently cancer-free). One detail of this implementation is that it does a linear interpolation between the age buckets specified in the table.
Example: If patient is age 34, we assume the that risk is half way between the risk, specified at age 29 and age 39.
To run tests with your current python interpreter:
python -m unittest discover -v .
To run Python 2.7/3.6 tests, first make sure you have tox installed. Then:
tox