Initial Project challenge:

In Varient’s mission to link people with rare diseases to treatment information for their specific condition, we discovered that the information gap among the rare disease community is, in part, caused by the lack of information standardization in the industry. With each hospital, clinic, and genetic testing agency documenting their data differently, one person’s genetic report can look miles apart from the other, making it inefficient to extract meaningful information.

This project was developed in Windows and Python 3.9.10, and thus Windows is required to run the scripts. Ensure Google Cloud is installed on the Windows Machine, initialized, and the API key JSON file is set as an enviroment variable

Then, setup the project enviroment

git clone https://github.com/not-cosmo/Geneticheck.git
cd Geneticheck
python -m venv venv
.\venv\Scripts\Activate.ps1
pip install -r requirements.txt

To run the Genetic Report Image Detection, simply run python geneDetection_gcloud.py. This script outputs the detection PASS/FAIL of the Gene in the filename and creates an image copy with a bouding box of all Gene instances. This script has a 100% detection accuracy for all sample images.

An opensource alternative was also written, however does not acheive the 100% accuracy the Google Cloud method achieves: python geneDetection_pytesseract.py

To run, simply run python generateGeneticReports.py to generate Genetic Reports for all stored gene mutations.

Alternatively, specific genes can be inputted as list of arguments. eg. python .\generateGeneticReports.py ACTC CFTR