Same Group, Different Spatial Stories: Hidden Variability Across Biological Replicates in Spatial Transcriptomics
CONFERENCE OF THE EUROPEAN SOCIETY FOR SPATIAL BIOLOGY, Barcelona, İspanya, 28 Eylül - 01 Ekim 2026, ss.1-2, (Özet Bildiri)
- Yayın Türü: Bildiri / Özet Bildiri
- Basıldığı Şehir: Barcelona
- Basıldığı Ülke: İspanya
- Sayfa Sayıları: ss.1-2
- Ankara Üniversitesi Adresli: Evet
Özet
Spatial transcriptomics technologies have enabled high-resolution characterization of tissue architecture and multicellular organization within intact tumors. However, biological interpretation of spatial datasets remains highly dependent on the specific tissue section selected for analysis. In heterogeneous solid tumors, independently acquired sections from the same biological condition may capture distinct ecological regions, potentially introducing substantial variability into downstream analyses. Despite increasing adoption of spatial transcriptomics platforms, the impact of section-to-section variability on biological interpretation and computational analysis remains insufficiently explored.
In this study, we investigated spatial variability between independently processed tumor sections derived from an orthotopic murine 4T1 breast cancer model using the 10x Genomics Xenium imaging-based spatial single-cell transcriptomics platform with a targeted 5,000-gene panel. Spatial analyses were performed on tissue sections obtained directly from primary tumor samples collected in vivo.Rather than focusing on treatment-associated differences, we specifically evaluated how two biologically matched naïve tumor slides generated divergent spatial and transcriptional outputs despite belonging to the same experimental group. Comparative analyses revealed notable differences in cellular composition, immune-cell localization, stromal organization, and neighborhood architecture between sections. Spatial distributions of lymphoid and myeloid populations varied substantially across slides, accompanied by differences in inferred cell–cell interaction patterns and regional tissue organization.Importantly, these discrepancies were observed despite consistent experimental processing and quality-control metrics, suggesting that tissue-section selection itself represents a major source of variability in spatial transcriptomic studies. The findings indicate that individual slides may preferentially capture distinct tumor ecological states, including immune-rich, stromal-dominant, or tumor-cell–dense regions, thereby influencing downstream biological conclusions. To address this challenge, we further evaluated analytical strategies that may reduce section-dependent variability, including expression normalization, integrated data harmonization, cell-type proportion scaling, neighborhood-aware analyses, and spatially constrained comparative frameworks. Our observations support the need for spatial transcriptomics pipelines that explicitly account for section-level heterogeneity rather than treating each slide as a fully representative snapshot of the entire tissue.
Collectively, this work highlights section-dependent variability as a critical but underrecognized factor in spatial transcriptomics and proposes a conceptual framework for improving robustness, reproducibility, and interpretation of spatially resolved tumor ecosystem analyses.