正文：
北京航空航天大学和清华大学的研究团队近日在Nature Communications期刊上发表了一项重要研究成果。该团队开发了一个名为soScope的统一生成框架，利用人工智能技术大幅提高了空间组学平台的数据质量和空间分辨率，相比原始分辨率提升了36倍。
空间组学技术是一种能够同时对组织和细胞的空间位置和分子组成进行分析的强大工具。然而，这些技术在应用中往往受到空间分辨率的限制，这阻碍了科学家们对复杂组织结构的深入研究。现有的计算方法主要集中在转录组学数据的分辨率提升上，缺乏对新兴空间组学技术的适应性。
soScope框架能够整合来自组学、空间关系和图像的多模态组织信息，并通过分布先验与组学特定建模联合推断出具有增强分辨率的组学谱。该框架不仅适用于传统的空间组学技术，如Visium、Xenium和spatial-CUT&Tag，而且能够扩展到新兴的多组学技术，如spatial-CITE-seq和空间ATAC-RNA-seq。
通过综合评估多种空间组学平台，soScope提高了对肠道和肾脏等组织结构的识别能力，揭示了在原始分辨率下无法解决的胚胎心脏结构，并纠正了测序和样本处理中出现的样本和技术偏差。
这项研究为空间组学技术的发展提供了新的方向，为科学家们提供了更强大的工具来探索复杂的生物组织结构，这对于理解疾病的发病机制、开发新的治疗方法以及推动生命科学领域的发展具有重要意义。
英语如下：
News Title: “AI Revolution: Beihang and Tsinghua Join Hands with Nature Journal to Boost Organ Profiling Resolution by 36 Times”
Keywords: High-resolution, AI Technology, Spatial Genomics
News Content:
Title: Team from Beihang and Tsinghua Significantly Enhance Spatial Genomics Resolution with AI
Article:
A research team from Beihang University and Tsinghua University has recently published an important research achievement in the Nature Communications journal. The team developed a unified generative framework named soScope, which utilizes artificial intelligence technology to significantly improve the data quality and spatial resolution of spatial genomics platforms, a 36-fold enhancement over the original resolution.
Spatial genomics technology is a powerful tool that allows for the simultaneous analysis of the spatial location and molecular composition of tissues and cells. However, these technologies often encounter limitations due to spatial resolution, which hinders scientists’ in-depth study of complex tissue structures. Existing computational methods primarily focus on enhancing the resolution of transcriptomics data, lacking adaptability to emerging spatial genomics technologies.
The soScope framework can integrate multimodal tissue information from genomics, spatial relationships, and images, and infer enhanced-resolution genomic signatures through joint distribution priors and genomics-specific modeling. This framework is not only suitable for traditional spatial genomics technologies such as Visium, Xenium, and spatial-CUT&Tag, but it can also be extended to emerging multimodal technologies like spatial-CITE-seq and spatial ATAC-RNA-seq.
By comprehensively evaluating various spatial genomics platforms, soScope has improved the identification of tissue structures such as the intestines and kidneys, revealing embryonic heart structures that were not apparent at the original resolution, and correcting sample and technical biases that arose during sequencing and sample processing.
This research provides a new direction for the development of spatial genomics technologies, offering scientists more powerful tools to explore complex biological tissue structures. This is of great significance for understanding the mechanisms of disease, developing new treatments, and advancing the field of life sciences.
【来源】https://www.jiqizhixin.com/articles/2024-08-14-10
Views: 1