腾讯发表最新Nature子刊论文：推出AI大模型，从单细胞转录组翻译单细胞蛋白质组

Core Viewpoint - The article discusses the development of a pre-trained large generative model called scTranslator, which translates single-cell transcriptomes to proteomes, addressing challenges in single-cell proteomics and enabling multi-omics data generation [3][5]. Group 1: Technology Overview - Single-cell proteomics faces challenges such as limited coverage, low throughput and sensitivity, batch effects, high costs, and strict experimental protocols [3]. - The scTranslator model is inspired by natural language processing and genetic translation processes, allowing for the inference of missing single-cell proteomes from single-cell transcriptomes [5]. Group 2: Validation and Applications - The research team conducted systematic benchmarking and validation of scTranslator across various profiling technologies (e.g., CITE-seq, spatial CITE-seq, REAP-seq, NEAT-seq), cell types (e.g., monocytes, macrophages, T cells, B cells), and tissues (e.g., blood, lung, brain) [5]. - scTranslator demonstrates accuracy, stability, and flexibility in a wide range of disease contexts, including infections, metabolism, and tumors [5]. - The model shows superiority in assisting various downstream analyses and applications, such as gene/protein interaction inference, perturbation prediction, cell clustering, batch correction, and cell origin identification in pan-cancer data [5].