Ma Huifang, Zhang Xingyuan, Qu Yilong, Zhang Anru R, Ji Zhicheng
Department of Biostatistics and Bioinformatics, Duke University School of Medicine, Durham, NC, USA.
Computational Biology and Bioinformatics Program, Duke University School of Medicine, Durham, NC, USA.
Genome Biol. 2025 Jun 18;26(1):173. doi: 10.1186/s13059-025-03648-w.
Spatial transcriptomics enables spatially resolved gene expression analysis, but accompanying histology images are often degraded by fiducial markers and background regions, hindering interpretation. To address this, we introduce Vispro, an end-to-end automated image processing tool optimized for 10× Visium data. Vispro includes modules for fiducial marker detection, image restoration, tissue region detection, and segmentation of disconnected tissue areas. By enhancing image quality, Vispro improves the accuracy and performance of downstream analyses, including tissue and cell segmentation, image registration, gene expression imputation guided by histological context, and spatial domain detection.
空间转录组学能够实现空间分辨基因表达分析，但附带的组织学图像常常因基准标记和背景区域而退化，妨碍解读。为解决这一问题，我们引入了Vispro，这是一款针对10× Visium数据进行优化的端到端自动化图像处理工具。Vispro包括用于基准标记检测、图像恢复、组织区域检测以及分离组织区域分割的模块。通过提高图像质量，Vispro提升了包括组织和细胞分割、图像配准、基于组织学背景的基因表达插补以及空间域检测等下游分析的准确性和性能。
