GeneTree

GeneTree Greeting Page
Example view of a combination of family triees

Inspiration

There is no available software for comparing multiple family trees and generating a likelihood percentage of multiple family trees overlapping each other

What it does

Just that, our tool takes multiple seemingly unrelated family trees and finds connections between them. It then generates a confidence level of how sure it is these trees overlap correctly.

How we built it

We used python to parse the data files and make relations between everyone in the reported family history. We then take other reported family history and try to find overlaps with variables such as Disease, Age of Onset, Age of Death, etc. We then take these results and display them in a user friendly webpage for analysts to view. Analysts can even upload their patients own family history and attempt find overlaps in other trees.

Challenges we ran into

This problem in itself is very complicated, even just trying to find ways to describe the relationship between family members programmatically proved to be difficult. Displaying these relations in a user friendly way also proved difficult.

Accomplishments that we're proud of

We can successfully map out relations across families, and we have an algorithm to get a confidence level for how sure we are two or more medical histories overlap.

What we learned

This is an extremely useful tool if done properly and could help a lot of people due to the issues of reported medical history. Tools like this don't currently exist for analysts or physicians and doing them by hand would take an immense amount of time. This is a good stepping stone in the right direction for medical history. On the programming side, we learned that there isn't really a great pre-existing way to depict a family tree, making it difficult just to get started. We learned just sticking to it and not being afraid to ask for help from mentors and peers will help you to overcome and task.

What's next for GeneTree

Next we'd like to get the visualization for the tool to be better. We'd like analysts to be able to interact with the family tree and more easily see just where the data overlaps to better understand the confidence.