A revolutionary case for a genetics and technology company that undertakes large scale animal breeding. The outcome of this project caused international excitement in the genetics community. Project lead Nicholas Hogan traveled to the US to present the case at the world’s biggest geneticists conference.
Can we speed up predictions for the effects of cross-breeding?
It can take generations to understand the full effects of crossbreeding. Additionally, the value of an animal has to be predicted at an early age so that it can still be reproduced.
We trained a machine to meet or exceed human predictions.
Using a deep learning model we were able to exceed the results of traditional models. Using DNA data and ultrasound scans we trained a deep learning model to make a phenotype prediction.
The time to predict dramatically decreased while accuracy increased.
The prediction time is now much shorter and invariant to the size of the animal pool, which allows the geneticists to make much faster iterations.