A new computational tool could change the way scientists judge whether lab-grown kidney tissue truly resembles the human organ—and whether it can accurately reveal the damage caused by drugs and environmental chemicals. Researchers Lee, Kang, Lim and colleagues have introduced KiGEP, a quantitative algorithm designed to calculate human kidney similarity and nephrotoxicity in human kidney organoids, according to a study published in Experimental & Molecular Medicine.
Kidney organoids are miniature, three-dimensional tissues grown from human stem cells. They contain populations of kidney-like cells and can reproduce selected features of renal development, filtration and drug response. Although these organoids have become a powerful alternative to conventional cell cultures and some animal experiments, one major problem has remained: researchers have often relied on visual inspection or a limited number of molecular markers to decide how closely an organoid resembles a real human kidney.
That uncertainty matters because kidney research is particularly sensitive to differences in cellular identity and organization. The human kidney contains specialized structures, including glomerular cells that participate in filtration, tubular epithelial cells that reclaim water and solutes, and supporting stromal and vascular populations. If an organoid contains the wrong balance of these cell types, or if its cells are immature, a substance may appear harmless—or dangerously toxic—for reasons that do not reflect what would happen in a human patient.
KiGEP is intended to address this challenge by turning the concept of “kidney likeness” into a measurable biological score. Rather than treating organoid quality as a simple yes-or-no judgment, the algorithm evaluates molecular and cellular information and compares it with reference features from human kidney tissue. In principle, this kind of analysis can integrate multiple layers of evidence, such as gene-expression patterns, the presence of kidney-specific cell populations and the relative abundance of markers associated with mature renal function.
The algorithm’s second purpose is to quantify nephrotoxicity, or kidney injury caused by a chemical or drug. Kidney toxicity is a major obstacle in pharmaceutical development. Many compounds fail during clinical testing because they damage renal tissue, while some toxic effects are difficult to detect in standard laboratory systems. A computational framework that connects organoid similarity with measurable injury signals could help investigators distinguish between a genuinely kidney-relevant response and an artifact produced by an immature or poorly characterized model.
This approach could also improve comparisons between laboratories. Organoids can vary substantially depending on the stem-cell line used, the growth medium, the timing of differentiation and the physical conditions in which they are cultured. Two research groups may describe their organoids as “kidney-like” while producing tissues with different cellular compositions. By applying the same quantitative framework, scientists could compare models more consistently and identify which organoid preparations are best suited for particular experiments.
The potential impact extends beyond drug screening. A reliable similarity score could support disease modeling, personalized medicine and studies of developmental kidney disorders. Researchers might use patient-derived cells to create organoids, then assess how closely those tissues reproduce the patient’s renal biology before testing therapies. KiGEP could also help identify whether a model is sufficiently mature for a specific experiment, reducing the risk of drawing conclusions from a tissue that only superficially resembles a kidney.
Yet an algorithm cannot eliminate every limitation of organoid biology. Kidney organoids generally lack the full architecture, blood circulation, immune interactions and mechanical forces of a functioning human kidney. Their filtration systems may be incomplete, and the timing of their development may not perfectly match human maturation. A high similarity score therefore should not be interpreted as proof that an organoid is a complete replacement for an actual organ. Instead, it can provide a more rigorous way to define what the model captures—and what it still misses.
The significance of KiGEP lies in this shift from visual confidence to quantitative validation. As organoid research expands, the field increasingly needs common standards for measuring cellular identity, maturity and toxic responses. By offering a computational strategy for calculating human kidney similarity alongside nephrotoxicity, the new framework could make kidney organoids more reproducible and more useful for biomedical research. If validated across diverse organoid systems and chemical exposures, KiGEP may help transform miniature kidneys from promising experimental models into more dependable tools for predicting human safety.
Subject of Research: Human kidney organoids, kidney similarity and nephrotoxicity assessment
Article Title: KiGEP: a quantitative algorithm for calculating human kidney similarity and nephrotoxicity in human kidney organoids
Article References: Lee, S., Kang, H.M., Lim, J.H. et al. “KiGEP: a quantitative algorithm for calculating human kidney similarity and nephrotoxicity in human kidney organoids.” Experimental & Molecular Medicine (2026). https://doi.org/10.1038/s12276-026-01785-1
Image Credits: AI Generated
DOI: 10.1038/s12276-026-01785-1
Keywords: KiGEP, human kidney organoids, nephrotoxicity, kidney similarity, quantitative algorithm, drug screening, stem-cell research, renal disease modeling
