In border collies (dog breed) exist idiopathic epilepsy (with an unknown cause or mechanism of apparently spontaneous origin). The model of it’s inheritance is yet unknown (supposedly poligenic). My idea was to get data of dogs suffering from the disease, add the most of the ancestry tree and calculate the probability the pup from a given mating will be affected.
I have some doubts whenever to pursue this thing because:
- the data about affected dogs may be a bit corrupted (some may be non idiopathic epilepsy, some may be other neurological disorder with similar symptoms -> false positives, also there is a case of a lot of false negatives when owner/breeder wants to hide it)
- from what I’ve heard genetical algorithms either work or they don’t and it’s hard to predict what will be the case and this one requires a lot of work on the side of data collection.
- it requires a ton of work
Has anyone worked on similar issue? Or has more experience that I do (none 
and knows whenever this could work?