Background: Human leukocyte antigen (HLA) complex molecules play an essential role in immune interactions by presenting peptides on the cell surface to T cells. With significant progress in deep learning, a series of neural network based models have been proposed and demonstrated with their good performances for peptide-HLA class I binding prediction. However, there still lack effective binding prediction models for HLA class II protein binding with peptides due to its inherent challenges. In this work, we present a novel sequence-based pan-specific neural network structure, DeepSeaPanII, for peptide-HLA class II binding prediction. Compared with existing pan-specific models, our model is an end-to-end neural network model without the need for pre- or post-processing on input samples. Results: The leave-one-allele-out cross validation and benchmark evaluation results show that our proposed network model achieved state-of-the-art performance in HLA-II peptide binding. Besides state-of-the-art performance in binding affinity prediction, DeepSeqPanII can also extract biological insight on the binding mechanism over the peptide and HLA sequences by its attention mechanism based binding core prediction capability. Conclusions: In this work, we present a novel neural network structure for peptide-HLA class II binding prediction. It has state-of-the-art performance and could display insightful information it learned benefiting from attention module we carefully designed. Without requiring additional data, this structure could be applied to other related sequence problems. The source code and trained models are freely available at https://github.com/pcpLiu/DeepSeqPanII.

Discovering novel magnetic materials is essential for advancing the spintronic technology with significant applications in data communication, data storage, quantum computing, and etc. While Density functional theory (DFT) has been widely used for designing materials, its high computational demand for estimating the magnetic ground states of even a single material limits its ability to explore the vast chemical design space for finding the right materials for spintronic applications. In this work, we developed a computational framework combining generative adversarial networks (GAN), machine learning (ML) classifiers, and DFT for de novo magnetic material discovery. We used the CubicGAN generative crystal structure design model for creating new ternary cubic structures. Machine learning classifiers were developed with around 90% accuracy to screen candidate ternary magnetic materials, which are then subject to DFT based stability validation. Our calculations discovered and confirmed that Na6TcO6, K6TcO6, and BaCuF6 are stable ferromagnetic compounds, while Rb6IrO6 is a stable antiferromagnetic material. Moreover, Na6TcO6 and BaCuF6 are found to be half metals that are highly favorable for spintronic applications. Due to the structural differences, A6MO6 materials have a higher thermal capacity (Cv) compared to BaCuF6. At 300 K temperature, Cv of A6MO6 materials is around 1100 J/K.mol and that of BaCuF6 is about 176 J/K.mol. This work demonstrates the promising potential of deep generative design for discovering novel functional materials.