Image Credit: GEMINI
Image Credit: GEMINI

Prologue

This analysis presents a systematic analysis of publicly available kidney spatial-transcriptomics data. The underlying experiments were conducted by other research groups, and the original experimental objectives are not assumed here. Instead, the analysis begins with the available data: 128 samples generated using five measurement technologies, spanning two species and a range of biological contexts, including development and disease.

The purpose of this analysis is to demonstrate how these data can be analyzed systematically to characterize kidney biology. The analysis is organized into ten steps, each addressing a specific analytical objective. Together, these steps examine the structure, quality, and biological content of the datasets while distinguishing findings that are well supported by the available evidence from interpretations that require additional validation.

The emphasis throughout is on extracting biologically informative patterns from the data while clearly defining the limitations of the analysis.

The chapters are:

  1. The Data at Hand: the studies, platforms, conditions, and the quality of every sample.

  2. Five Platforms, One Language: how I made five incompatible technologies comparable, and what comparability cost.

  3. The Single-Cell Lens: what 1.9 million single cells say about kidney composition, on Xenium and CosMx.

  4. The Spot and Bin Lens: whole-transcriptome context from Visium and Visium HD, and how to unmix it.

  5. The Region Lens: compartment biology from region-level GeoMx profiles with no coordinates at all.

  6. The Architecture of the Kidney: spatial domains and the cortex-to-medulla program.

  7. Injury and Repair: the ischemia-reperfusion story, read at single-cell resolution.

  8. Disease Readouts: DKD, transplant rejection, and vasculitis through a spatial lens.

  9. The Spatially Variable Program: the genes that define where biology happens.

  10. Synthesis: what the analysis established, and what remains beyond even all five platforms.

Every quantitative result presented in this analysis is derived from the analyzed data, and each substantive claim is supported by a corresponding table or figure. Where the available data do not provide sufficient evidence for a particular conclusion, this limitation is explicitly noted. Analytical methods that do not perform adequately are also documented rather than omitted. Throughout, the analysis is guided by the biological questions of interest, with methodological choices serving those questions and the limitations of the data and methods treated as an integral part of the analytical process.

⚠️ Independent Research Disclaimer & Scientific Limitations

Declaration of Independence: The author declares no conflicts of interest or shared affiliations with the authors of the original studies whose public data are re-analyzed here. The computational re-analysis and results presented herein are completely independent and have no official connection to any of the original publications. The original studies are referenced solely for experimental design, data structure, and biological context.

Educational & Non-Clinical Purpose: This analysis is intended solely for bioinformatics educational and demonstration purposes. The datasets span human and mouse kidney tissue across development and disease. Under no circumstances should these findings be interpreted as clinical advice, used to guide patient care, or used as experimental suggestions for therapeutic interventions in patients.

Data Scope & Limitations: All analyzed data were retrieved from the public Gene Expression Omnibus (GEO). This analysis is provided “as-is,” does not seek orthogonal experimental validation, and should be interpreted accordingly.