Jizhong Lou

The paper identifies CD3ε as a primary trigger subunit in TCR activation through phase separation with Lck, leading to the design of TRIG molecules that activate T and NK cells via endogenous receptor signaling.

The paper identifies the Marangoni effect as a mechanism for intracellular transport of biomolecular condensates using C. elegans zygotes and in vitro experiments.

This study shows that PD-1 disrupts CAR-LCK condensates by binding CAR, and adding a CD3ε module stabilizes these condensates, enhancing CAR T cell activity against solid tumors.

This study reveals that PD-1 remodels SHP2 dynamics to inhibit T cell activation through a non-catalytic mechanism, independent of its phosphatase activity.

This study shows that PD-1 ligand interactions are force-dependent and soluble PD-L1 can block PD-1, suggesting new approaches for cancer immunotherapy.

This study shows CD28 co-stimulation forms signaling condensates with Lck, which PD-1 modulates by binding and disrupting these structures, leading to engineered CAR T cells with improved antitumor activity.

TCRAD designs antigen-specific TCR CDR3β sequences using deep learning, achieving high accuracy in binder discrimination and structure prediction, with experimental validation showing 17.2% functional candidates.

The paper introduces UMImap, a pipeline for accurate UMI identification and error correction in nanopore sequencing, enabling reliable full-length transcriptome analysis and discovery of novel isoforms.

GCRTcall improves RNA basecalling accuracy using a Transformer with gated convolution and relative position embedding, achieving state-of-the-art performance.

This paper introduces BaseNet, a transformer-based toolkit for nanopore sequencing signal decoding, showing improved accuracy through autoregressive and non-autoregressive decoding mechanisms.

The paper develops a deep learning-based tool, OPT-sgRNA, to design CRISPR sgRNAs that balance on-target activity and off-target effects using public CRISPR screen data.