Research interests

Our research is dedicated to developing computational algorithms and conducting large-scale data mining from single-cell multiomic data. We aim to unravel the intricate mechanisms of gene regulation, cellular crosstalk, and metabolic reprogramming that define cell identities. We have created methods and web resources specifically designed for single-cell multiomics, spatial multi-omics, single-cell CRISPR screen data analysis, signal enhancement, and integration. By seamlessly integrating single-cell datasets with spatial and multimodal resolution, our objective is to establish connections between changes in cell identity and disease phenotypes in various biological systems, including cancer, development, and aging.

Developing intelligent algorithms for single-cell and spatial multi-omics

Single-cell and spatial multi-omic technologies have revolutionized our understanding of cellular heterogeneity in complex biological systems. However, corresponding analyses currently face various challenges, including inadequate resolution, coverage, and difficulty integrating and generating heterogeneous multi-modal data. To address these issues, we have developed a series of intelligent algorithms. STRIDE can improve the resolution of Visium-like ST. Cellist enables accurate cell segmentation in high-resolution ST datasets. SCRIP and SCRIPro constructs gene regulation networks (GRNs) using single-cell and spatial multiomic data. EvaCCI evaluates cell-cell interaction algorithms, and SCREE analyzes multi-modal single-cell CRISPR screening data. These algorithms have enhanced the resolution of single-cell spatial omics data and provided systematic solutions to challenges such as single-cell multiomic integration, signal enhancement, and the construction of cell-fate-specific GRNs.

Modeling cell identities using generative AI and large-scale multi-modal data

Cell identities in a multicellular system are regulated by both intrinsic factors, including gene regulation, and extrinsic factors, such as cellular crosstalks. We have demonstrated the tight connection between intrinsic epigenetic regulations and cell-fate determination in mouse IVF and SCNT embryo development (Nature, 2016; Nat. Cell Biol., 2018; Cell Stem Cell, 2018, 2022; Cell Res., 2022). Currently, we are developing generative AI models pretrained on large-scale single-cell multi-modal datasets. These models aim to uncover the collaborative effects between gene regulation, cellular crosstalk, and other environmental factors, such as metabolites and mechanics, on cell identity regulation. As a pioneering work, we have introduced SELINA, which utilizes a multi-adversarial domain adaptation network to automatically annotate cell types with a large-scale human scRNA-seq reference pretrained. We hope to use generative cell identity models to provide valuable insights into the mechanisms driving cell identity transitions and to further guide and remodel the transition process.

Discovering the diversity and plasticity of cell identities in the tumor immune microenvironment

Cancer arises from the evasion of immune surveillance, and the immunosuppressive tumor microenvironment has a strong impact on tumor development and therapy resistance. Our objective is to integrate single-cell and spatial multi-omics data with comprehensive bioinformatics data analysis to investigate the effects of intrinsic gene regulation and extrinsic environmental factors on altering immune cell identities in the tumor microenvironment (TME). Moreover, we aim to develop potential methods for remodeling the TME in cancer treatment. As a preliminary step, we have developed a comprehensive scRNA-seq data resource TISCH for analyzing gene expression and cell-type composition in the TME. Additionally, we have constructed a pan-cancer cell identity tabula TabulaTIME and discovered widespread profibrotic ecotypes that regulate tumor immunity. Currently, we are collaborating closely with oncologists and immunologists to study the mechanisms of TME evolution and immunotherapy resistance in different types of cancer (Cell 2024, Genome Med., 2023, and EMBO J., 2023).