Background

This reporistory was done as part of a research project for genetic matching. It supposed to improve the problem of finding the best implicated genetic haplotypes when the data of some relatives is known and the sample has too many ambiguities. Currently I believe that better solutions exist using hidden markov models with machine learning on actual data. Additionally, as genetic sequencing is now improving, and of much lower cost than the past, there are much less ambiguities in the genetic samples and such solutions are not needed. The algorithmic solution is still valuable as in the view of privacy it may be benficial to use much less "identifying" genetic sequencing methods.

The project requires confidential genetic files to run. Therefore, it will probably be impossible to run.

Running & Dependencies

To run this project you will need these libraries:

1. openpyxl
2. jupyter notebook (jupyter)
3. matplotlib

Install with pip, for windows users you could use (working for python 3.5+, be sure that py command recognized, if not you will need to change PATH variable):

py -m ensurepip
py -m pip install --upgrade pip
py -m pip install openpyxl
py -m pip install jupyter
py -m pip install matplotlib

Be sure that project's genetic files configured in main.py, you will need 4 files:

1. The genetic data file, currently supported is excel file format
2. The ambiguities file
3. The haplotypes file
4. The frequency file each line corresponding to the line in the haplotype file that has the same number

Run this project with python 3.6 by running main.py:

py main.py

Also PyPy recommended (it's reducing running time, install here: https://pypy.org/download.html).

After that you can see the results (solution.csv file by default) using jupyter notebook (there is a notebook file inside notebook folder) and after some configuration.