ID #1073 SpotQC: spatial-aware pre-analytic quality controls for spatial transcriptomics
Brent Orr, Sariah Allen, Julie Jeschke, Nesha Dorsey, Christina Tucker, Zahangir Alom, Yiran Li, Louis El Khoury, David Ellison, Paul Northcott, Selene KooAbstract
Background
Spatial transcriptomics (ST) assays have become popular methods to gain insights regarding the distribution and relationship between cell types within tissue. Current methods, including 10X Xenium and Visium HD assays, are costly and have relatively high failure rates in archived formalin-fixed paraffin-embedded (FFPE) tissue from clinical samples. Existing quality control (QC) metrics, such as DV200, evaluate quality from bulk extracted RNA but do not capture regional heterogeneity of RNA quality across a tissue section. Thus, these QC metrics may misrepresent the quality of the annotated region for ST and allow no mechanism to direct annotations to the highest quality regions. Cost-effective and spatially aware QC metrics are therefore needed to improve the success rate of ST assays.
Methods
RNA in situ hybridization (ISH) assays were developed using RNAscope and hybridization chain reaction (HCR) to recognize housekeeping genes with low, medium, and high expression levels across multiple tissue types. RNA-ISH was quantified via machine-supervised scoring and visualized as QC intensity plots for RNA target signals across tissue sections. The method was applied to a cohort of brain and non-brain solid tumors that had undergone ST by 10X Xenium or Visium HD assays. ST success was correlated with QC metrics from traditional methods and RNA ISH.
Results
ST success on clinical samples showed minimal correlation with traditional QC metrics. Strong correlation was observed between RNA ISH-based QC metrics in the annotated region and ST success. Probes targeting medium expression genes showed the highest concordance in our tumor cohort, with 4 to 6 spots/cell representing the minimum cutoff to optimize ST success.
Conclusions
We developed and validated a cost-effective RNA ISH-based method to ascertain regional quality in FFPE sections for ST. These methods should improve future success rates in ST projects performed on archived clinical samples.