Supplementary Figure 4: Schematic illustration of PrimateAI, the deep learning network architecture for pathogenicity classification.
From: Predicting the clinical impact of human mutation with deep neural networks
The inputs to the model include 51 amino acids (AA) of flanking sequence for both the reference sequence and the sequence with the variant substituted in, conservation represented by three 51-AA-length position-weighted matrices from primate, mammal, and vertebrate alignments, and the outputs of networks for training secondary structure and solvent accessibility (also 51 AA in length).
