This image depicts diseased human liver tissue. Tissue stained in red shows fibrosis around the central vein. Tissue stained in green shows the bile ducts. Tissue stained in yellow shows immune cells. A white line is drawn around a region of interest on the image.

NanoString GeoMx Whole Transcriptome Atlas (WTA)

The GeoMx Digital Spatial Profiler (DSP) platform provides whole-transcriptome profiling of human or mouse tissue using barcoded, photocleavable probes. This approach involves identifying regions of interest through morphological markers (up to three antibodies plus a nuclei stain). It's a sequencing-based platform that is compatible with formalin-fixed paraffin embedded (FFPE), fresh frozen and fixed frozen samples.

The GeoMx DSP enables tissue morphology-driven, spatially resolved (within 1 μm), high-plex digital profiling of hundreds of proteins and tens of thousands of messenger RNA (mRNA) directly in FFPE or fresh-frozen human and rodent tissue sections. This technology combines standard immunofluorescence and in situ hybridization techniques with oligonucleotide barcoding quantification, allowing for the multiplexed detection of proteins and mRNAs in a single sample.

The multiplexing detection can range from up to 96 proteins or several tens of transcripts if using the nCounter device for the readout or hundreds of proteins and the whole transcriptome if quantified using the next-generation sequencing (NGS) readout. Applied to two serial sections, simultaneous (mRNA) and protein profiling of the same sample is possible. Furthermore, the addition of the Immuno-Oncology Proteome Atlas enables the integrated proteogenomic profiling of FFPE or fresh frozen tissue sections where the whole transcriptome and over 570 proteins are collectively targeted. This allows for high-plex, multimodal omic interrogation in the same slide.

The GeoMx DSP technology involves histological, information-based and user-directed liberation of ultraviolet (UV)-photocleavable oligonucleotide barcodes, their quantification by nCounter or NGS techniques, and mapping of the tag counts to the original tissue location. This results in a spatially resolved digital profile of protein or mRNA abundance.

Here's how it works:

  • 5 µm to 10 μm FFPE or fresh frozen sections are stained with up to three fluorescent antibodies to identify tissue landmarks and cell types. Multiple sections or tissue microarrays mounted on up to four slides can be analyzed together.
  • From about 20 to over 570 target proteins are reacted with oligonucleotide-tagged antibodies.
  • Target mRNAs ranging from several tens included in focused panels to the whole transcriptome are hybridized in situ with barcoded probes.

In addition to mouse and human immuno-oncology-focused panels and neuroscience-focused panels and whole-transcriptome atlases, custom protein and mRNA probes can be included in the experiments.

After generating high-quality images, the user selects multiple regions of interest of up to 660 µm-by-785 μm within which multiple subareas can be defined by:

  • Geometric shapes.
  • Contour lines.
  • A grid pattern.
  • Immunofluorescent staining-based, software-aided segmentation of tissue areas, such as tumor versus stroma.
  • Target cell types, such as immune versus parenchymal cells.

UV exposure of the tissue within these masks liberates the photocleavable barcodes, which are collected by microcapillaries into individual wells of 96-well plates. These oligonucleotides can be quantified by using standard NanoString nCounter or Illumina sequencing (100 paired-end and 50-base reads per μm2 collection area). Barcode or sequencing read counts can be pooled from multiple sources representing cells, groups, replicates or patients for downstream analysis, as appropriate.

GeoMx DSP comes with built-in data analysis software. The data analysis workflow includes quality control and normalization operations. The intuitive user interface supports several data visualization and higher-order analysis methods, including clustering, dendrograms, relating expression levels to tissue architecture, differential protein or gene expression analysis, statistical evaluation, and pathway analysis.

Here's how it works:

  • The sequencing data (FASTQ files) are processed with the GeoMx NGS Pipeline.
  • After sequencing, reads are trimmed, merged and aligned to a list of indexing oligonucleotides to identify the source probe. The unique molecular identifier region of each read is used to remove polymerase chain reaction duplicates and duplicate reads, thus converting reads into digital counts.
  • Raw Illumina counts are upper quartile (Q3) normalized using the GeoMx software and standardized quality control threshold settings, as recommended by the manufacturer. These recommendations are a gene detection rate for segments at 5% to 20% and a segment detection rate at 5% to 25%.
  • Low-quality samples are removed from analysis if the geometric mean of aligned reads from all probes is less than the geometric mean of the negative control probes. They also are removed if the Q3 of the counts in each area of interest is less than the geometric mean of the negative control probes in the data.

Statistical analyses of the data generated are performed using R (version 4.1.2). The expression of individual genes is determined as a log2 fold change. A linear model is used to fit the RNA count data. The Benjamini and Hochberg method is used to control false discovery rate for multiple testing. The analysis is carried out using lme4 package (version 1.1.31) and stats package (version 4.1.2) with emmeans (version 1.8.4.1).