At #AGBT19 there is a palpable movement from dissociated single cells to single cells in situ
Every year at the Advances of Genome Biology and Technology you can see what the future holds. When long-read sequencing was introduced in 2011, people here at AGBT19 are still talking about that Pacific Biosciences fireworks display on the beach here at Marco Island. The presentation that afternoon stated boldly that a 15 minute, whole genome sequence was only five years away.
Here in 2019 several talks are about long reads from PacBio and from Oxford Nanopore, and their usefulness for structural variant detection in addition to de novo genome assemblies.
Another wave of innovation
When 10X Genomics was introduced here in 2015, another wave of innovation was set to take place looking at RNA expression at the single-cell level rather than as a bulk sample. Cellular Research was getting started (to be acquired by BD Life Sciences), MissionBio started showing some posters (if I remember correctly there was a 6-plex qPCR poster a UCSF cell-sorting core facility director presented at AACR in 2016), and also in 2016 Berkeley Lights presented an interesting poster at AGBT.
And here in 2019 many, many talks are about scRNA-seq. Last night when chatting with a friend about The Broad’s Aviv Regev’s opening talk the first full day of AGBT 2019, “Cell atlases as road maps to human disease”, he joked (somewhat seriously I might add) that she could have covered only one of the three topics she presented in her 30-minute talk, and she had to skip the fourth topic due to lack of time. That particular presentation was memorable in that it was hard to keep up with the speed of information being shared.
It reminded me of talks that Chris Mason gives (Weill Cornell), who as one of the conference organizers is not speaking at this year’s AGBT, but his (preliminary) twin study in space work will be presented by a NASA scientist later this morning by John Charles, NASA Johnson Space Center (retired). “From here to Mars: How the Twins Study and the year-long ISS mission have moved us closer to the red planet”.
Spatial genomics is here
For those who are curious, the topic Dr. Regev didn’t get to was called ‘DNA microscopy’, and seeks to look at spatial information in two dimensions.
Although imaging of cells and tissues has been a cornerstone of biology ever since cells were discovered under the light microscope centuries ago, a gap has emerged between these methods and genomic measurements. Although both forms of measurement characterize a single biological reality, they profile the microscopic world differently. Microscopy in itself illuminates spatial detail, but does not capture genetic information unless it is performed in tandem with separate genetic assays. Conversely, genomic and transcriptomic sequencing do not inherently capture spatial details.From Weinstein and Zhang et al. DNA microscopy: Optics-free spatio-genetic imaging by a stand-alone chemical reaction. doi: https://doi.org/10.1101/471219
In brief, DNA microscopy uses a clever use of Unique Molecular Identifiers (UMI’s), reverse transcription in-situ, and amplification of overhang primers for concatenation in-situ to get proximity information in two-dimensional space.
In a clinical diagnostics context there is an inherent trade-off between imaging analysis that a pathologist would use and the high multiplexing molecular profiling techniques such as genome-wide copy-number variation interrogated with a clinical microarray. This contrast is illustrated below (figure borrowed from this Nanostring GeoMx DSP brochure).
A brief review of Nanostring’s GeoMx Digital Spatial Profiler (DSP) at #AGBT19
Last year I wrote up this review of the Nanostring GeoMx DSP, at that time called simply the Digital Spatial Profiler. In brief the system uses either oligo probes (for mRNA interrogation) or antibodies labeled with oligos (for multiplex protein analysis) and these oligos have a UV-photocleavable linker. Using the microscope as part of the system, ROIs (Regions Of Interest) are selected as geometric shapes, or by certain fluorescent morphology markers (up to four can be used).
Using digital micromirror technology, UV light will cleave the oligo labels in a specific region, and a microcapillary will aspirate the liquid without touching the sample. This non-destructive interrogation means that the slide can be stored and later reprobed again if necessary. The specific region’s labels are deposited into a 96-well plate, hybridized automatically to NanoString barcodes and read out quantitatively using their nCounter technology.
NanoString had a ton of content at #AGBT19 and the conference isn’t even at the half-way mark. Their presentations include the following: a four-hour “Pre-AGBT Spatial Genomics Summit” (agenda is online here) that included several alternative technologies and approaches, a sponsored lunchtime workshop with two invited speakers, and several posters for both the technology and the application of this technology by early-access partners. In addition last night Katherine McNamara of Stanford University gave an interesting talk in the Concurrent Cancer session called “Spatial and temporal profiling of protein and RNA in the tumor-immune microenvironment during short-term targeted therapy in HER2-positive breast cancer using GeoMx™ Digital Spatial Profiler”.
NanoString plans to launch the GeoMx DSP at the AACR Conference in Atlanta Georgia, March 29 – April 3 2019, and have already stated their first-year launch goal of selling 30 instruments has already been met. With data demonstrating up to 60-plex protein assays and over 1,400 gene targets, the ability to re-interrogate archival samples ‘is a game-changer’ according to one speaker. Priced at about $300,000, there is clearly an appetite for such capability.
10X Genomics is not standing still
Three months ago 10X Genomics acquired Spatial Transcriptomics in December of 2018, and for those who are not familiar with the technology they use a slide with poly-adenylated probes attached to produce cDNA in two-dimensions. Here is Spatial Transcriptomics 2016 Science reference, and they had just started to have a commercial presence; at last October’s ASHG Spatial Transcriptomics had a booth presence.
10X Genomics said their offering which they call Visium Spatial platform and will be produced as a collection of protocols, reagents, slides and software. They expect it to be available in the second half of 2019. They did not specify in their workshop how much better in spatial resolution it will be over the original Spatial Transcriptomics offering (the original platform had 100 um spots spaced 200 um apart). According to Sonja Vickovic (currently at the Broad Institute), this contains from 3 to 30 cells (depending on the cell type) and the geometry can interrogate about 20% of the available surface.
A key limitation to the technology, however, is its ability to interrogate only fresh-frozen samples. This is not a minor point, as the ability to bring liquid nitrogen into the operating room is no trivial task for flash-freezing tissue samples, rather than dropping the tissue sample into formalin.
In one of the conference talks, “Spatial mapping of host-microbiome interactions in the gut” Dr. Vickovic mentioned they are going to beads in hexagonal (densely-packed) wells, with 2 um features which may give a clue as to what the final Visium product may look like.
These are early days for spatial genomics
One key application driver for spatial genomics (and proteomics) is research on therapy response to immunotherapy. “Hot” and “Cold” tumor microenvironments (TME) is a complex topic, with a wide variety of immune cell-types overlaying the tumor tissue, and an active area of research.