To enable personalized genetics and medicine, it is important yet highly challenging to accurately predict disease-causing mutations from the sequences alone at high throughput. To meet this challenge, we build upon recent progress in machine learning, ne