Xenium Renal Cell Carcinoma

We generated Xenium data from four clear-cell renal cell carcinoma (ccRCC) patient samples each consisting of tumor and adjacent tissue. A customized gene panel was used consisting of 280 genes provided by a base panel (10X's initial breast cancer gene panel) and 100 genes selected to target kidney-specific genes, T-cell subsets, and genes that were previously observed to show significant variation in expression in previously generated single-cell RNA-Seq data.This dataset was used it part to demonstrate an improved cell segmentation method <b>Proseg</b>, in the paper <i>Cell Simulation as Cell Segmentation</i>.Further details:The RCC FFPE blocks were provided by Northwest Biotrust, under a NWBiospecimens protocol, Seattle. NW BioTrust, a core service for patient consenting, and NWBioSpecimen, a core service for procurement and annotation of research biospecimens. The analysis was performed according to the IRB file/approval number NHS #6007-1061. No age or gender information is available. Informed written consent was obtained from each subject or each subject’s guardian.<br>Xenium allows customization of gene panels, which we took advantage of for this project. We selected 100 genes to add on to a standard breast cancer base panel of 280 genes. Fourteen of these were manually selected from prior interest and the remaining 86 were selected using an existing in an unsupervised manner using a prior scRNA-Seq dataset. Briefly, multinomial logistic regression was fit to cell type labels in this data and genes were prioritized by their absolute regression coefficient, as an approximation of how useful the gene is in distinguishing known cell types. The 100 add on genes were finalized in collaboration with 10X, removing some with very high expression risking optical crowding.We followed the Demonstrated Protocol Xenium in Situ Tissue Preparation Guide (CG000578 Rev A) to place samples on Xenium slides (10.45mm x 22.45mm). Deparaffinization and decrosslinking steps followed protocol CG000580 (Rev A). The Xenium In Situ Gene Expression User Guide (GC000582 Rev A) was used for the remaining workflow. The human breast base panel (280 genes) and a 100 gene custom add-on panel (See supplementary) were combined at the probe hybridization step. The slides were loaded in the Xenium Analyzer as directed by CG000584 (Rev A). Imaging was conducted in cycles on the instrument, capturing data across multiple Z-planes and fluorescence channels to build a spatial map of transcripts within selected scanned regions. For two samples we selected five regions within each tissue representing tumor and adjacent based on histology. We scanned the entire tissue as a single region for the second pair samples. The onboard analysis pipeline included image processing, nucleus and cell segmentation, RCA product image registration, transcript decoding, quality scoring, and deduplication. Quality scores were estimated using controls for calibration, ensuring the accuracy of the final output data. After decoding, duplicate transcripts were reconciled, and a cell-feature matrix was generated, allowing for further analysis and integration with existing datasets.

我们从4例透明细胞肾细胞癌（clear-cell renal cell carcinoma, ccRCC）患者样本中获取了Xenium空间转录组数据，每例样本均包含肿瘤组织与癌旁组织。本研究使用的定制化基因面板包含280个基因：其中280个基因来自基础面板（10X初始乳腺癌基因面板），另有100个基因为靶向筛选所得，涵盖肾脏特异性基因、T细胞亚群相关基因，以及此前已报道的在单细胞RNA测序（single-cell RNA-Seq, scRNA-Seq）数据中存在显著表达差异的基因。
本数据集部分用于验证一款改进的细胞分割方法**Proseg**，相关研究发表于论文《Cell Simulation as Cell Segmentation》。
详细信息如下：本研究中的肾细胞癌福尔马林固定石蜡包埋（Formalin-Fixed Paraffin-Embedded, FFPE）组织块由西雅图的Northwest Biotrust按照NWBiospecimens协议提供。其中，NW BioTrust为患者知情同意提供核心服务，NWBioSpecimen则负责科研生物样本的获取与注释工作。本研究的分析工作依据伦理审查委员会（Institutional Review Board, IRB）审批文件编号NHS #6007-1061开展。研究未获取受试者的年龄与性别信息，所有受试者或其监护人已签署书面知情同意书。
Xenium平台支持定制化基因面板，本研究充分利用了这一特性。我们在280个基因的标准乳腺癌基础面板基础上，新增了100个基因。其中14个基因基于前期研究兴趣手动筛选，剩余86个基因则通过对已有单细胞RNA测序（scRNA-Seq）数据集采用无监督方式筛选得到。简言之，我们对该数据中的细胞类型标签拟合多项逻辑回归模型，并以回归系数的绝对值对基因进行优先级排序，以此近似衡量该基因在区分已知细胞类型中的应用价值。新增的100个基因最终与10X团队共同确认，剔除了部分表达量极高、可能引发光学拥挤的基因。
我们按照《Xenium原位组织制备操作指南（CG000578 Rev A）》的示范流程，将样本置于规格为10.45mm×22.45mm的Xenium玻片上。脱蜡与解交联步骤遵循CG000580（Rev A）操作流程。后续实验流程均参照《Xenium原位基因表达用户指南（GC000582 Rev A）》执行。在探针杂交步骤中，我们将280个基因的人源乳腺癌基础面板与100个基因的定制化新增面板（详见补充材料）进行混合。按照CG000584（Rev A）的指引，将玻片置于Xenium分析仪中。仪器通过多轮循环完成成像，采集多个Z平面与荧光通道的数据，以构建选定扫描区域内转录本的空间分布图谱。
针对其中2例样本，我们基于组织学特征在肿瘤与癌旁组织中各选取5个区域；对于另外2例样本，我们将整个组织作为单个区域进行扫描。机载分析流程涵盖图像处理、细胞核与细胞分割、滚环扩增（Rolling Circle Amplification, RCA）产物图像配准、转录本解码、质量评分与重复转录本去重等步骤。质量评分通过校准对照进行估算，以保障最终输出数据的准确性。解码完成后，我们对重复转录本进行校正，并生成细胞-特征矩阵，可用于后续分析以及与已有数据集的整合。