Multi-omics enables a systematic comprehension of information transmission across various levels. The progress in bulk and single-cell sequencing technologies, along with associated computational approaches for multi-omics, has greatly supported the growth of systems biology and precision medicine. The multi-omics technologies and analyses have been widely applied in neuroscience studies, ranging from the detection of alterations at molecular scale and their localization in different brain regions. Single-cell sequencing technologies have revolutionized our perspective on the fundamental constituents of the brain, reshaping how we examine the cellular foundations. These technologies dismantle intricate biological systems into individual cellular elements, offering an ultra-high-resolution exploration of brain cell populations. This approach also facilitates the delineation of cell subgroups possessing distinct functional attributes, inference of regulatory mechanisms shaping cell identities and cell type-specific reactions, and establishment of connections to brain development and diseases.  

1. Integrating Multi-omics Data with Novel Methods 

Multi-omics data often have a characteristic where the number of features is much greater than the number of samples. Effectively integrating complex multidimensional data to extract meaningful information is a key and hot topic in the field of bioinformatics. In our laboratory, we have developed new strategies for integrating multi-omics data, including single-cell transcriptomics, chromatin accessibility, spatial transcriptomics, and proteomics, by leveraging deep learning models. These new methods are of great significance for gaining a deeper understanding of the molecular mechanisms of diseases, identifying new molecular biomarkers, and understanding the relationship between the molecular basis and phenotypes. 

2. Brain Cell Atlas Research Based on Single-Cell Multi-omics Data 

The brain consists of various cell types that are interconnected to form specific neural circuits supporting cognitive and behavioral functions. The advancement of single-cell multi-omics methods enables us to comprehensively understand the cellular and molecular composition of the brain, thus unraveling the molecular basis of complex brain functions. By integrating data from various single-cell omics approaches, we employ different analytical strategies to unveil the cellular composition of the primate brain, comprehend the molecular underpinnings of cell and molecular diversity throughout evolution, and identify molecular biomarkers in the primate brain during disease, development, and aging processes through computational analysis. 

3. Development of Gene Editing Tools Based on Deep Learning 

The efficiency and simplicity of the CRISPR-Cas system have greatly promoted its applications across various fields in life sciences. However, the challenge remains in selecting the optimal editor and sgRNA. We address this challenge by constructing deep learning models to learn and predict the editing efficiency and off-target effects of various gene editing tools. This allows us to in silico select the best guide RNAs for in vivo experiments. Additionally, we leverage omics data to develop computational workflows for the evaluation, enhancement, and creation of a new generation of highly efficient, persistent, and safe gene editing tools for disease therapy.