Interpretation of non-coding genome remains an unsolved challenge in human genetics due to impracticality of exhaustively annotating biochemically active elements in all conditions. Deep learning based computational approaches emerge recently to help interpret non-coding regions. Here, we present LOGO (Language of Genome), a self-attention based contextualized pre-trained language model containing only two self-attention layers with 1 million parameters as a substantially light architecture that applies self-supervision techniques to learn bidirectional representations of the unlabelled human reference genome. LOGO is then fine-tuned for sequence labelling task, and further extended to variant prioritization task via a special input encoding scheme of alternative alleles followed by adding a convolutional module. Experiments show that LOGO achieves 15% absolute improvement for promoter identification and up to 4.5% absolute improvement for enhancer-promoter interaction prediction. LOGO exhibits state-of-the-art multi-task predictive power on thousands of chromatin features with only 3% parameterization benchmarking against the fully supervised model, DeepSEA and 1% parameterization against a recent BERT-based DNA language model. For allelic-effect prediction, locality introduced by one dimensional convolution shows improved sensitivity and specificity for prioritizing non-coding variants associated with human diseases. In addition, we apply LOGO to interpret type 2 diabetes (T2D) GWAS signals and infer underlying regulatory mechanisms. We make a conceptual analogy between natural language and human genome and demonstrate LOGO is an accurate, fast, scalable, and robust framework to interpret non-coding regions for global sequence labeling as well as for variant prioritization at base-resolution.
T cell receptor-engineered T cells (TCR-Ts) have emerged as potent cancer immunotherapies. While most research focused on classical cytotoxic CD8+ T cells, the application of CD4+ T cells in adoptive T cell therapy has gained much interest recently. However, the cytotoxic mechanisms of CD4+ TCR-Ts have not been fully revealed. In this study, we obtained an MHC class I-restricted MART-127-35-specific TCR sequence based on the single-cell V(D)J sequencing technology, and constructed MART-127-35-specific CD4+ TCR-Ts and CD8+ TCR-Ts. The antitumor effects of CD4+ TCR-Ts were comparable to those of CD8+ TCR-Ts in vitro and in vivo. To delineate the killing mechanisms of cytotoxic CD4+ TCR-Ts, we performed single-cell RNA sequencing and found that classical granule-dependent and independent cytolytic pathways were commonly used in CD4+ and CD8+ TCR-Ts, while high expression of LTA and various costimulatory receptors were unique features for cytotoxic CD4+ TCR-Ts. Further signaling pathway analysis revealed that transcription factors Runx3 and Blimp1/Tbx21 were crucial for the development and killing function of cytotoxic CD4+ T cells. Taken together, we report the antitumor effects and multifaceted killing mechanisms of CD4+ TCR-Ts, and also indicate that MHC class I-restricted CD4+ TCR-Ts could serve as potential adoptive T cell therapies.
Three patients with nasopharyngeal carcinoma developed binaural secretory otitis media 12, 2, and 0.5 years after radiotherapy, respectively. The secretions subsided after conventional drug and drainage treatments, but hearing continued to deteriorate until severe loss was documented in both ears. After examination of the eardrum and tympanum, patients were enrolled in a clinical trial in the first half of 2019 (ClinicalTrials.gov: NCT03818347). After 0.5, 1 and 2 months of continuous hydrogen–oxygen therapy, our first three patients reported different levels of improvement in binaural hearing. This is the first report to show that, after treatment for nasopharyngeal carcinoma, hearing loss can be alleviated using hydrogen–oxygen therapy.