Purpose: By preserving tissue architecture, spatial transcriptomics offers new insights into the biological processes occurring within native micro-environments. To further advance our understanding of complex biology, there is a need for a high-throughput platform that combines a large capture area, unbiased whole-transcriptome detection, and cellular resolution. We have re-engineered our flow cell technology to create the Illumina Spatial Solution (ISS), an advanced NGS-based platform that pushes the boundaries of spatial transcriptomics. The Illumina substrate uses ~1 µm features along with the ability to capture thousands of unique transcripts per cell and allows for flexible tissue placement in a large 7.5 cm2 active area. With an upfront histological imaging step, ISS uses an AI-based nuclei segmentation model trained on comprehensive datasets for accurate nuclei identification, anisotropic expansion and segmentation. This segmentation approach enables precise cell binning, cell typing and marker gene identification built in addition to scalable DRAGEN™ technology. With intuitive tools for image processing, data analysis, and spatial visualization, Illumina Spatial Image Tool, DRAGENTM and Illumina Connected Multiomics software tools streamline secondary and tertiary analysis.

Methods: We applied ISS to fresh frozen adult human cortex and mapped spatial RNA expression in neuronal and non-neuronal cells. By sequencing spatially barcoded cDNA on a NovaSeq X instrument, we detected ~4000 unique transcripts and ~2000 genes per cell, as well as identifying >45,000 unique genes in the sample. Spatial maps of specific gene expression revealed the expected layering of human cortex, correct expression of marker genes for excitatory and inhibitory neurons, as well as for glia and vascular cells. Our advanced image registration enabled high-precision alignment of transcript data with H&E imaging, facilitating accurate cell binning. Leiden clustering of cell-binned data produced maps of distinct cell types that could then be further classified by assigned marker gene expression.