What data do we have, what are we predicting and why are we doing that?
From raw ChIP-seq data to model input.
Documentation for our dataset and neural network architecture API.
Training BPNet to predict the ChIP-Nexus profiles and total counts for all four TFs.
Comparing different number of dilated convolutional layers, different number of channels in each dilated convolutational layer, and different kernel sizes for the first layer.