...architectures—CNNs, recurrent neural networks (specifically, long short-term memory, LSTM), and dCNNs, for high-resolution chromatin accessibility prediction.
Building on this, we incorporate ConvNeXt V2 (Woo et al. 2023) blocks, originally developed for vision tasks, as feature extractors across all...