EVE is a model for the prediction of clinical significance of human variants based on sequences of diverse organisms across evolution. It uses fully unsupervised deep learning trained on amino acid sequences of over 140K species. We make predictions for all single amino acid variants of disease related genes, which come in the form of a score that ranges from 1, most pathogenic, to 0, most benign.
This project has been developed by:
Jonathan Frazer*, Pascal Notin*, Mafalda Dias*, Aidan Gomez, Joseph K. Min, Kelly Brock, Yarin Gal** and Debora Marks**
* these authors contributed equally
** corresponding authors
Marks Lab - Harvard Medical School
OATML - Oxford Applied and Theoretical Machine Learning Group
You can read EVE’s publication
here.
You can download the EVE scores, ClinVar data, gnomAD frequencies and more for all proteins modeled so far
here.
Alternatively you can explore the results for a particular protein by using the search bar above.
To create additional EVE models yourself or to simply browse through the code, check out our
repository on GitHub!