Fig. 5: DeepNull DNN model architecture. Each rectangle represents one layer and all layers are fully connected.
From: DeepNull models non-linear covariate effects to improve phenotypic prediction and association power
Shaded layers use the ReLU activation and the non-shaded layers do not use an activation function (i.e. linear connection). The input is the set of known covariates and the output is the predicted phenotype.
