Expert Perspectives on Spatial Transcriptomics

eBook Perspectives from Global Pioneers in Spatial 10x Genomics Perspectives from Global Pioneers in Spatial 2 Charting biology: The case for spatial transcriptomics In multicellular organisms, no single cell exists in a vacuum. Each cell is influenced by the others in its surrounding environment, the signals that pass between them determining their functions. In fact, the maintenance of cellular and tissue spatiality—or lack thereof—carries significant implications for everything from Alzheimer’s disease and cancer to autoimmune conditions and immune responses to infection. Spatial biology provides unprecedented insights into this organization by shedding light on global cell-type organization, cell–cell interactions, heterogeneous spatial niches, novel gene programs, critical ligand– receptor signaling networks, and biomarkers associated with disease. Examining biological molecules within appropriate morphological context has been possible since the 1940s. While early publications for immunohistochemistry1 and in situ hybridization2,3 proved spatial profiling was possible, they were limited in the number of detectable targets and resolution. Today, we can map hundreds of transcripts up to the whole transcriptome4–7 while achieving single cell or even subcellular resolution. Spatial transcriptomic technologies can be divided into two primary categories: imaging-based and sequencingbased. Imaging-based techniques, such as Xenium Spatial assays, label and image transcripts directly within a tissue slice. Sequencing-based techniques, such as Visium Spatial assays, release and capture transcripts or ligation products corresponding to transcripts from tissues and then analyze those products using an external sequencer. Which method is right for you, ultimately, depends on your specific experimental questions and/or needs. This eBook introduces you to six researchers at the forefront of adopting spatial biology in their fields of oncology, immunology, and neuroscience. They share how taking a spatial approach has allowed them to make incredible breakthroughs, discuss what they considered when choosing a spatial platform, and give advice to researchers starting to explore spatial transcriptomics. “Architecture is what ultimately distinguishes a living cell from a soup of the chemicals of which it is composed. How cells generate, maintain, and reproduce their spatial organization is central to any understanding of the living state.” Whellan DJ and O’Connor CM. Function follows form. JACC Heart Fail 9: 482–483 (2021). doi: 10.1016/j.jchf.2021.02.012 FPO 10x Genomics Perspectives from Global Pioneers in Spatial 3 Table of contents Mapping immunity: Dr. Reina-Campos discusses the power of spatial transcriptomics 04 From allergies to autoimmunity: Dr. Poholek’s guide to cutting-edge spatial research 08 Spatial biology at scale: Dr. Banovich on decoding disease with a new dimension of context 13 Defining tumor boundaries: Dr. Ye’s spatial exploration of immune escape mechanisms 17 Shaping future therapeutic possibilities: Dr. Freytag on the power of spatial insights in clinical research 20 Unveiling tumor secrets: Dr. Watson’s spatial multiomics approach to glioblastoma research 26 10x Genomics Perspectives from Global Pioneers in Spatial 4 Miguel Reina-Campos, PhD Assistant Professor La Jolla Institute for Immunology Mapping immunity: Dr. Reina-Campos discusses the power of spatial transcriptomics Tell us about why spatial transcriptomics is so important for the research that you do. Absolutely. For decades, immunologists have been studying how the different pieces of the immune system work. They’ve looked at these pieces using many different technologies and applications. We understand a great deal about how these different pieces work, their diversity, and how they function, but much less is known about the specifics of where those species have to operate in tissues. It has always been a Dr. Reina-Campos’ research focuses on specialized immune cells called tissueresident T cells that live deep within our organs, such as the colon or the liver, to provide fast and robust protection against future infections and tumors. In particular, his lab is interested in unlocking the mechanisms that make tissue-resident T cells proficient at their jobs in hopes of revealing new approaches to revamp the immune response against cancer. 10x Genomics Perspectives from Global Pioneers in Spatial 5 mystery to really observe those immune cells within their natural environments. My lab is studying how to better leverage the immune system that exists within tissues. Funnily enough, we had never been able to observe those cells in the tissues they are protecting. Prior to spatial transcriptomics, our observations had always been very indirect. We either had to make smoothies out of the tissues to separate the cells to study them, but obviously losing their natural environment, or we would do histological stains but at a very low plex of maybe 3–5 markers. However, that in itself is not enough to understand the very rich ecosystem these cells live in. We also study the differentiation of the immune system, which means looking at what genes the immune cells are expressing and when they are expressing them. We have been using single cell technologies like scRNA-seq to look at their RNAs, proteomics to look at their proteins, and even metabolomics to look at the metabolites. But, we had never been able to put the two things together—where these cells are in a tissue and what they are expressing. With spatial transcriptomics, for the first time, we were able to observe our favorite immune cells in unperturbed, natural environments at a resolution that we had never been able to look at before and at a plex that we had never even dreamed of before. These spatial transcriptomics technologies, I know lots of folks are very interested in them, but for us, they came as a necessity. We needed this technology to get to the answers that we were looking for. Do you have some specific examples of how spatial has impacted the line of research that you’re looking at? The cells that we study are mighty and powerful, but they are not very abundant. We needed something that would enable us to look at those small lymphocytes, sometimes located next to bigger and much more abundant tissue cells, and see what they express. Using Xenium technology has enabled lots of cool research coming out of our lab, and other labs, in defining the The image depicts three overlapping layers of biological information in the mouse small intestine—protein (left), to cell types (middle), to individual transcripts (right)—offering a new window into the secret life of tissue-resident memory T cells (purple, right). Image credit: Reina-Campos lab. 10x Genomics Perspectives from Global Pioneers in Spatial 6 spatial patterning of how the immune system is interconnected within a tissue architecture. For example, we’ve looked at the small intestine and found that these immune cells, which come in different flavors—some are more effectors while some are more progenitors—are actually localized in different parts of the intestine. This spatial distinction of immune subtypes is something that we had never been able to look at before. Not only that, because we have spatial context, we can now understand what signals are driving that heterogeneity and better understand how tissue–immune networks happen in the context of infection and tumors. You talked about needing to see cells that weren’t very abundant. What other considerations did you take into account when you were choosing what type of spatial you were going to pursue? At the beginning, when these new technologies came along and there were different emerging players in the field of spatial transcriptomics, we were technology agnostic as long as we could get to the bottom of our research. We just wanted to get good data and by that I mean getting the maximum sensitivity and the maximum surface area for the maximum amount of transcripts that we could. But then we realized that having robust technology is pretty important and were disappointed with other platforms. When we started using 10x Xenium, something that we really liked was that it was very robust. It was an almost end-to-end solution where you could have your tissues, start in a prep, and then end up looking at your data a couple of weeks later, already getting some insights into your problem. Other technologies did not have that end-to-end solution yet. We appreciated having that already built in. Nowadays, being reproducible and robust is a big deal for us. Another consideration was that it works with formalinfixed paraffin-embedded (FFPE) samples. I think that’s what most people have access to, and it unlocks thousands of samples coming from clinical practice. Other technologies struggled with FFPE, and we understood the chemistry that 10x was providing could work very reliably with FFPE. So, robustness and FFPE compatibility were the two considerations that were defining for us in choosing 10x over others. I ultimately think having multiple options will be best for us researchers. “With spatial transcriptomics, for the first time, we were able to observe our favorite immune cells in unperturbed, natural environments at a resolution that we had never been able to look at before and at a plex that we had never even dreamed of before.” Spilling the (immune) secrets of our guts Our intestines play a crucial role in providing protection from infections thanks in part to the presence of tissue-resident memory CD8 T (TRM) cells. To understand the two different types of TRM cells present in the intestines and how their location impacts the role they play, Reina-Campos et al. harnessed the high resolution of Xenium single cell spatial imaging and the whole transcriptome profiling power of Visium sequencing-based spatial transcriptomics. This powerful approach, alongside a novel spatial CRISPR knockout experiment, revealed how fundamentally connected a T cell’s location is to dictating its functional state. Read the publication to discover what cellular secrets were revealed > 10x Genomics Perspectives from Global Pioneers in Spatial 7 A lot of people see the promise of spatial biology, but may be hesitant to jump into a new technology. What advice would you give for people who are just starting out in spatial to help them be successful? I think you kind of have to address the elephant in the room, which is the price. People who get stuck on the price without looking at the data often change their minds once they see the data and all the information that can be unlocked. My advice is to run one pilot experiment to see the data, so you can really appreciate what it can do for your project. But still, price remains a challenge and a gatekeeper for many. So moving forward, we should aim to encourage companies to have a lower level of entry and democratize this technology so more people can use it, because it really is an amazing technology. It’s my hope that very soon it will cost a tenth of the price so more people can use it. Are there any new directions or capabilities within the field of spatial biology and spatial technologies that you would be really excited to see pop up in the future? Absolutely. One of the considerations for choosing 10x Xenium over other technologies is its non-destructive approach to spatial transcriptomics. Meaning that when you’re done with your experiment, you still have your tissue mostly intact, which enables you to pair it with multiomic readouts. Having 5,000 genes and up to a couple dozen protein stains you can maybe look at after your experiment, that’s one thing. Other omics modalities like metabolomics would be very powerful. I anticipate people will want to know all the genes, but the downside is sensitivity challenges at higher plexy. I’m hoping in the future we get that right balance of higher sensitivity and higher plex. Is there anything else you want to highlight, a particular publication from your lab, or any other thoughts you want to share on spatial? Spatial transcriptomics requires a blend of expertise that no single person can possibly have. I was super lucky to be part of a team, first with a fantastic mentor that enabled this research. I’ve had amazing colleagues supporting different aspects of our projects, all of which are required for the spatial transcriptomics to succeed. From the bioinformatics to the sample preparation to the histology to the in vivo animal experiments, you really need a 2–4 person team to do any of these big experiments. I want to acknowledge that it’s not a oneman show. These are complex experiments that require full teams. Because of the work of our team, we were able to find that the heterogeneity of the immune system is spatially imprinted, at least in the gut, and we suspect in many other tissues as well. Those findings were accepted for publication three weeks ago and will soon be published in Nature. We’re very excited that more and more people are going to be using this. I can’t wait to see where it leads us. 10x Genomics Perspectives from Global Pioneers in Spatial 8 From allergies to autoimmunity: Dr. Poholek’s guide to cutting-edge spatial research What was your introduction to spatial transcriptomics, and why did you decide that it would be necessary for your research? First, I’m the director of the Health Sciences Sequencing Core at the University of Pittsburgh, which provides Next Generation Sequencing services for the institution. We are always looking at new technologies in the sequencing space that we think might be in demand for researchers at the university and to make sure that we’re able to offer cutting-edge approaches to answer their research questions. Amanda Poholek, PhD Assistant Professor, Department of Immunology Director, Health Sciences Sequencing Core University of Pittsburgh Dr. Poholek is an assistant professor at the University of Pittsburgh. Her lab aims to understand how immune cells sense different signals and tissue environments to decide what kind of cells they need to become. Of particular interest is cell response to signals that should cause activation but, due to failure to tolerate non-harmful stimuli, lead to autoimmune and allergic diseases. Additionally, Dr. Poholek serves as the Director of the Health Sciences Sequencing Core. 10x Genomics Perspectives from Global Pioneers in Spatial 9 In my own research, we had an unexpected finding that the transcriptional repressor Blimp-1 was required for the differentiation of T helper 2 (Th2) immune cells specifically in the lung. These are the cells that can drive the physiologic symptoms of allergic asthma, such as mucus production, airway hyperresponsiveness, and smooth muscle constriction. Prior to our work, although we knew a lot of the factors critical for the development of Th2 cells, we hadn’t appreciated that Blimp-1 was required for that process. It was really a surprise because other work in other systems where Th2 cells develop, such as responses to worm antigens, had not demonstrated that Blimp was important for promoting an effective response but, rather, for constraining an effective response. Based on these disparate functions of Blimp-1 in different contexts, we decided to directly test if the route by which you experience the allergen actually drives a difference in the genetic pathways that you need to form these pathogenic Th2 cells driving allergic asthma. And it turned out that there was a difference. If you intranasally administer an allergen to an animal, then you need Blimp-1 to become a Th2 cell. However, if you administer the same allergen subcutaneously or systemically, you can form Th2 cells without needing Blimp-1 at all. These data really suggested that there was a tissuespecific pathway through the lung, driving an effector T-cell response that we had not previously understood. I H&E VisiumHD H&E overlay 8μm bins Visium data from mediastinal lymph node three days following house dust mite exposure. H&E stain (left) and graph-based clustering of FFPE Visium HD (right). Inset shows a HEV by H&E (left), overlay with transcriptional clustering (middle) and 8-μm spatial bins (right) for analysis. Image credit: Poholek lab. 10x Genomics Perspectives from Global Pioneers in Spatial 10 wanted to understand this at the initiation of the response to better understand how a naïve T cell interprets the information it receives during activation in the context of the tissue that experiences the antigen. Since the lymph node is the place where the T cells typically see antigen for the first time, we wondered if, by looking at the same exact response to the same exact antigen in lymph nodes draining two different tissue sites, we would reveal information about the pathways and spatial niches needed for those immune cells to become the effector cells that drive allergic asthma. I proposed in a grant to use spatial transcriptomics as an unbiased mechanism to examine these two different lymph nodes, draining either the skin or the lung. When you look at the initiation of the immune response to an allergen, how do those two sites differ? We are still working to answer this question but it’s what got us interested in spatial transcriptomics as a technology. While setting up these systems, we made an observation in the lung-draining lymph node that was so striking to us that we paused there and dived in. This work was published in Nature Immunology recently, where we identified spatial microniches of the cytokine IL-2 that were important for the formation of Th2 cells in allergic asthma through Blimp-1. We used spatial transcriptomics to identify Th2 cells in the lymph node, and then applied a new algorithm developed by our very talented collaborator, Dr. Jishnu Das, called SLIDE, which stands for Significant Latent Factor Interaction Discovery and Exploration. SLIDE can identify latent factors or hidden variables in highdimensional datasets that underlie specific observable features. Dr. Das was eager to use SLIDE on a spatial dataset, and our localization of Th2 cells in the lymph node provided an ideal opportunity. SLIDE identified IL-2Ra in a significant latent factor that was supporting differences in the localization of Th2 cells in the lymph node. Then, we were able to validate that IL-2 was acting as a local microniche in the lymph node to promote initiation of Blimp-1 to drive Th2 cells using other methods. I think our example of how we got into spatial transcriptomics is very typical—you have a biological question that requires spatial tissue context but also highlights the need for new computational approaches that can reveal biological insight. There are different flavors of spatial transcriptomics these days, imaging-based and then NGS-based systems. Which type are you using for your work, and what considerations did you take into account when deciding for your specific project? Currently, we’re using both. We started with NGS-based because those were the ones that were commercially available. Specifically, we used the 10x Genomics Visium Spatial Gene Expression and Curio Biosciences Seeker™ platforms, and that’s what we used in our recent publication. We’re now working with both the Visium and imaging-based Xenium platforms. There are several important considerations when thinking about these two different flavors. Now that we have higher resolution with Visium HD, the sequencingbased applications offer truly unbiased approaches at the single cell level and the ability to stain the same tissue section for imaging that you use for spatial transcriptomics. I think there’s some real power there if you aren’t 100% sure what you’re looking for, so it’s really useful in a discovery setting. Imaging-based platforms are also incredibly powerful at the single cell level. And now that they have probe sets that have gotten up to 5,000–6,000, you can do a lot of discovery-based work there, too. However, “I think our example of how we got into spatial transcriptomics is very typical—you have a biological question that requires spatial tissue context but also highlights the need for new computational approaches that can reveal biological insight.” 10x Genomics Perspectives from Global Pioneers in Spatial 11 imaging-based platforms are more beneficial if you have a really targeted set of genes you’re interested in. You can do more high-throughput studies, get more tissue into a run, and get a lot more information at a broader scale. Whereas for sequencing-based platforms, you get more information at the single cell level from the whole transcriptome but are limited in the amount of tissue that you can look at per run. As a core director, what advice do you give to people looking to get started in spatial but are unsure where to begin? The first piece of advice I’d give is the same for any sequencing application: ask yourself what biological question you’re trying to answer and then use the platforms or technologies that are the most appropriate to your question. Having a sense of what you’re hoping to find is really important when approaching any of these technologies. The second piece of advice is to really know your tissue. Our experience running a core and offering this as a service has allowed us to run over 30 projects and easily more than a dozen different tissues in mouse and human. Every tissue is different, and there are critical considerations when dealing with different tissue types. I advise talking to someone with broad expertise across multiple tissues if available at your institution. If not, take the time to do pilot runs with your tissue to make sure that it adheres the way you expect it to, that it stains the way that you expect it to, and that you’re able to get good quality sequencing data out of your tissue. We always do QC to know the RNA quality of our tissues before we proceed. If the quality of the RNA is not high, the returns are quite diminishing. It’s worth the effort to source and have very good quality material. If it’s material that’s easy for you to get, like from a mouse study, make every effort to collect that tissue with spatial in mind. You’re better off generating a new tissue block where you’ve thought clearly about RNA integrity rather than digging through the –80°C freezer to find something. Thinking into the future, what new directions or capabilities would you be particularly excited to see emerge in the arena of spatial biology? We get asked by many people how much protein staining can be done. Of course, people can do immunofluorescence, but it’s somewhat limiting. I think some of the barcoded antibody protein panels that could be offered commercially to go along with some of these platforms would be very useful for many projects. Along those same lines, I’m excited about adding additional modalities where you can do not just gene transcription but also protein and potentially mass spectrometry to look at metabolites or adding in epigenetics. Really, multimodal spatial is something that I’m excited to see come out. The other thing that we’ve done a little bit of, but, again, would be heavily utilized if more commercially available, would be spatial TCR and BCR sequencing for those of us in the immune space. “The first piece of advice I’d give is the same for any sequencing application: ask yourself what biological question you’re trying to answer and then use the platforms or technologies that are the most appropriate to your question.” 10x Genomics Perspectives from Global Pioneers in Spatial 12 Is there anything else you’d like to cover or highlight? Yes. As I mentioned when describing our recent findings in allergic asthma, it is important to highlight the need for innovative computational methods for the spatial space. We saw this a lot with single cell early on, where there was a lot of effort working to properly analyze and visualize single cell datasets. We’re in that same area right now for spatial. I think people are utilizing the single cell algorithms, sometimes effectively, but sometimes less effectively, for spatial. What we absolutely need, and we’re seeing this from some groups, are gold standard computational approaches for spatial that will win out in terms of their utility—approaches thinking about the data as spatial data rather than just trying to co-opt single cell algorithms into spatial algorithms. That’s an area for tremendous development that will also allow us to gain the most biological insight from these amazing new technologies in spatial transcriptomics. “What we absolutely need, and we’re seeing this from some groups, are gold standard computational approaches for spatial that will win out in terms of their utility—approaches thinking about the data as spatial data rather than just trying to co-opt single cell algorithms into spatial algorithms. That’s an area for tremendous development that will also allow us to gain the most biological insight from these amazing new technologies in spatial transcriptomics.” Inhaled, mapped, understood: Spatial microniches driving allergic asthma Our immune systems respond to lots of antigens, sometimes in very different tissue environments. Compelled to understand lung-specific responses to inhaled allergens—given that asthma is one of the most common and most costly diseases—He et al. combined Visium Spatial Gene Expression data with immunofluorescence to map the molecular pathways activated over time. This approach highlighted the unexpected but important role interleukin 2–mediated spatial microniches within the lung draining lymph node play in allergic asthma. Dive into these findings featured in Nature Immunology > 10x Genomics Perspectives from Global Pioneers in Spatial 13 Nicholas Banovich, PhD Director of the Center for Spatial Multi-Omics The Translational Genomics Research Institute (TGen), part of City of Hope Dr. Banovich’s research group leverages advanced genomic technologies, such as single cell RNA sequencing and spatial transcriptomics, to understand how genetic and molecular variations contribute to disease. This work aims to identify new biomarkers and therapeutic targets to improve patient outcomes across multiple disease areas, including chronic lung disease and cancer. Spatial biology at scale: Dr. Banovich on decoding disease with a new dimension of context Can you describe your research and your lab at TGen? My primary function is running a research lab. We are focused on understanding how gene regulatory changes impact disease outcomes—whether that’s initiation, treatment response, or disease progression. One of our major projects is focused on pulmonary fibrosis (PF). We also focus on correlative analyses from patients undergoing CAR T-cell therapies for brain tumors. So we use very different biological systems, but they’re unified by the set of tools and approaches. 10x Genomics Perspectives from Global Pioneers in Spatial 14 Behind the breath: Spatial insights into lung remodeling and PF According to the Pulmonary Fibrosis Research Foundation, each year approximately 50,000 new cases of PF will be diagnosed. Vannan et al. used Xenium single cell spatial imaging to map the structural and cellular changes that are a hallmark of idiopathic PF. This spatial gene expression of 1.6 million cells from 35 unique lungs allowed the team to explore the progressive remodeling of distinct spatial niches throughout the course of PF disease progression. Check out this groundbreaking research that earned a cover feature in Nature Genetics > We’ve built a foundation on single cell transcriptomics and using the single cell information to understand molecular dysregulation and disease at cell-type resolution. Over the past couple of years, we’ve started to put more of a focus on spatial technologies, and we’ve been using Xenium since February of 2023. My other role at TGen is operating and directing the Center for Spatial Multi-Omics (COSMO) technology core. At the core, we offer both Xenium and Visium HD assays as a service to internal TGen investigators and external researchers as well. Can you tell us more about how spatial has impacted your research projects? We think about the same types of questions that we thought about with single cell approaches, but spatial data has a lot of added benefits. It gives us understanding not just of the cell type–level dysregulation, but also the organization of cells into architectural niches. We can start asking: what types of cells are grouping together as disease progresses or in response to treatment? Do we see more local or broader changes in both the cell-type and molecular composition of tissues? Another really important benefit is not having to put your tissue sample through a dissociation procedure. For example, when dealing with the lung, there are a lot of really fragile cell types, which has its challenges for single cell experiments. In the alveoli—the little air sacs that fill up with air and move oxygen out of the air into your blood—there are two key cell types: alveolar type I (AT1) epithelial cells and capillary cells. The AT1 cells touch the air, and the capillary cells sit up against the AT1 cells. They’re both very thin and really integrated across the alveolus in three-dimensional space. Because of that, we’re prone to lose those with single cell approaches since they don’t survive the dissociation process. Some of these key cell types for understanding lung biology tend to be underrepresented in our single cell assays, but they perform quite well in the spatial data because we don’t have to go through that tissue dissociation step. When you started exploring spatial technologies, what did you take into consideration before choosing the right one for your lab? There were certainly spatial platforms out pre-Xenium, but none of them offered single cell resolution. Because of the complexity of the tissues we’re dealing with, particularly the complexity of the lung, we need that granularity to truly understand the biology. When we could look at the spatial component and really have that cell-type resolution, this is when spatial platforms became exciting to me. One of the things we really liked about Xenium, as we were looking at the available offerings, was that a lot of our tissues are formalin-fixed paraffin-embedded (FFPE) tissue, and Xenium has the ability to capture 10x Genomics Perspectives from Global Pioneers in Spatial 15 transcripts even from fairly old, degraded FFPE samples. These are samples that people collected, never with the intention of using them for a molecular assay. An anecdote that I like to share is that the oldest sample we’ve profiled was about 15 years old. It had just sat on my collaborator’s desk in a shoebox that he touched and dug through with bare hands over the course of those years—everything that just makes us, as RNA people, die on the inside, you know? Except, we were still able to get good data from it [with Xenium]. The high throughput of the Xenium system was also a key factor because we were really excited about the FFPE samples being amenable to tissue microarray (TMA) approaches. That we could have almost 5 cm2 of area per run started to help us envision how we might be able to do these spatial projects at scale. What I mean is, instead of just ten or fifteen samples, we could start thinking about hundreds of samples that we could process. Switching to your core director role, how do you explain spatial to new users? There are two user groups that we deal with. One is people already doing single cell who want to move to spatial. For them, it is really about highlighting that we still get that sort of single cell resolution, but with spatial context. The other user group we have are researchers who haven’t really been involved with single cell approaches. They may have done some sort of RNA in situ hybridization in the past. Here we really talk about this as a platform that allows you to take RNA in situ hybridization and multiplex this across hundreds or thousands of targets instead of just 1–4 targets. And, actually, I think in some ways spatial data is more intuitive for users who haven’t been doing genomics than single cell data is. This is because you can provide the files for the 10x Xenium Explorer, and people can go on and look at it like you would an immunohistochemistry experiment. You can click on and off the genes that you care about, and it’s very visually simple to understand that this blue dot corresponds to the gene I care about. Is there any advice you have for people who are just getting started with spatial? There are complexities on both the front end and the back end that affect new users. It certainly affected us as we started dealing with these platforms. We didn’t have histology experience before we started doing spatial transcriptomics. We were very comfortable with traditional genomics workflows like library prep and PCR. Tissue embedding, sectioning, and staining were all things we had to learn. So, if you’re coming from a genomics lab, there is a learning curve to how you process tissues. It’s important to either build those skills within your group, outsource to a core, or work with a histology group that has this experience. On the other hand, there is a ton of complexity around data analysis, and some of it is just a sheer numbers problem. Our first two Xenium runs far surpassed the number of cells we analyzed for the entirety of the Human Cell Atlas work for the lung. We went from “An anecdote that I like to share is that the oldest sample we’ve profiled was about 15 years old. It had just sat on my collaborator’s desk in a shoebox that he touched and dug through with bare hands over the course of those years—everything that just makes us, as RNA people, die on the inside, you know? Except, we were still able to get good data from it [with Xenium].” 10x Genomics Perspectives from Global Pioneers in Spatial 16 thinking about datasets that have a 2–4 million cell cap to a 60 million cell dataset. Ultimately, we had to move the analyses over to GPUs to handle the processing. The rest of the challenges, I think, are the fun challenges like: how do we really utilize this relational information? How do we link this to things we’re seeing in the histology images? How do you see spatial evolving, either in your research or for the wider field in general? This is a great platform, and I suspect this will become a commodity over the next couple of years. As development continues at 10x, we want to continue to be able to provide services around the newest technologies. We’re willing to go in and push on these technologies and figure out what works well and doesn’t work well. Then we want to share that knowledge with the community, so that our early mistakes and triumphs guide new users as they’re starting to set up. As people get better and better at using Xenium, the things we are doing now as front-edge users, like the large-scale TMA analyses, will probably be easily done by more researchers. I envision, and hope, that we’ll be helping to spearhead additional spatial applications on the horizon. Is there anything that’s really cool or interesting that you want to tell us about that you haven’t gotten a chance to? I think we’ve covered it! There’s all sorts of cool biology we’re looking at, but every person’s cool biology is going to be different, right? So the most exciting thing to me is to have more people thinking about these studies with bigger populations and really digging into their own biological questions with this kind of new dimension of information. Tissue sample showcasing the diverse histopathology of the PF lung (H&E stain; leftmost panel), including an airway, multiple blood vessels, remodeled epithelium, and fibrosis. A custom Xenium panel was used to profile 343 genes (second panel in). Individual nuclei and cells were viewed using DAPI and cellbound stains (third panel in). Individual features of interest were extracted using FICTURE (bottom right; Si et al., 2024 Nature Methods). This is part of a set of 45 lung samples profiled in Vannan, Lyu et al. (2025) in Nature Genetics; raw data available at Gene Expression Omnibus under accession GSE276945. Image credit: Annika Vannan, Arianna Williams-Katek, Nicholas Banovich. 10x Genomics Perspectives from Global Pioneers in Spatial 17 Defining tumor boundaries: Dr. Ye’s spatial exploration of immune escape mechanisms Tell us about your lab’s research focus. Using single cell and spatial omics data as the foundation of our analyses, we integrate innovative bioinformatics methods, basic experiments, and clinical research to assess how the tumor microenvironment (TME) responds to immunotherapy. We’ve developed novel computational methods for single cell spatial omics, such as Cottrazm, to tackle challenges in spatial tumor analysis. This allows us to explore how the TME drives the dynamic spatiotemporal evolution of tumor progression and Youqiong Ye, PhD Principal Investigator Shanghai Institute of Immunology Shanghai Jiao Tong University School of Medicine Dr. Ye is a principal investigator at the Shanghai Institute of Immunology. Her research focuses on how the spatial tumor microenvironment influences immune escape mechanisms in response to immunotherapy. A primary goal of Dr. Ye’s team is to identify novel therapeutic strategies targeting the tumor boundary microenvironment. 10x Genomics Perspectives from Global Pioneers in Spatial 18 influences therapeutic efficacy, with particular attention to the role of tumor boundary components in immune escape and immunotherapy effectiveness. This comprehensive approach enables us to better understand the interplay between the TME and therapeutic outcomes. Why did you pursue spatial transcriptomics for your research? Although single cell transcriptomics has revolutionized biological and medical research, a major limitation of scRNA-seq is its inability to capture spatial information. Spatial transcriptomics (ST) technologies have addressed this gap by enabling gene expression quantification in intact tissues while preserving the spatial localization of transcripts and the threedimensional organization of molecular processes. In our studies, we primarily focus on the spatial distribution of cells, cellular interactions, and network regulation within the tumor microenvironment and its response to immunotherapy. Therefore, the spatial positioning of transcripts is crucial, making ST an indispensable part of our research. How has spatial impacted your research? We first used ST in our research in 2021 after we discovered that SPP1+ macrophages and FAP+ fibroblasts were significantly enriched in colorectal cancer (CRC) tissues compared to adjacent normal tissues. At that time, the reviewers for our publication asked us to provide spatial evidence of the localization of these cell types, prompting us to use the 10x Visium platform. We found that both SPP1+ macrophages and FAP+ fibroblasts were located at the tumor boundary, where they formed an immune-excluded desmoplastic barrier—a dense barrier made of fibrous connective tissue—restricting T-cell infiltration and thus contributing to immune evasion in colorectal cancer. Since many of our studies use ST data, we have even developed the computational tool Cottrazm for spatial tumor analysis. The tool integrates ST data, morphological information from hematoxylin and eosin histological images, and single cell transcriptomics. These data are then used to delineate tumor boundaries and cell “spots” in tumor tissue, then leveraged to reveal cell type–specific gene expression signatures. More recently, we’ve also used tumor ST data to analyze pan-cancer microenvironment features. This project has focused on localizing functional enrichment of differentially expressed genes to specific tumor structures, teasing apart the TME’s cellular composition, identifying cell type–specific gene expression signatures, and characterizing cell–cell interactions. From this, we have developed a userfriendly database called SpatialTME that allows users to search for, visualize, and download results. Many of our ongoing studies continue to rely on ST, which has provided crucial information and accelerated our research progress. Mapping a new path forward for immunotherapeutic targeting of colorectal cancers What makes some tumors responsive to immunotherapy but others not? Qi et al. were able to correlate shorter progression-free survival in CRC with the presence of two distinct cell types— FAP+ fibroblasts and SPP1+ macrophages—inferring these cell types could be working synergistically. Visium sequencing-based analysis revealed that these two cell types were colocalized in the TME, and these areas were enriched for pathways associated with extracellular matrix and collagen fibril organization. This spatial insight was crucial to revealing how FAP+ fibroblasts and SPP1+ macrophages may work together to promote an immune-restrictive TME, earmarking them as attractive targets for future therapeutic strategies that can improve immunotherapy responsiveness. Explore the findings that made this one of Nature Communications’ top 25 health science articles of 2022 > 10x Genomics Perspectives from Global Pioneers in Spatial 19 What considerations did you take into account when choosing the type of spatial analysis you would use? The choice of which spatial technology to use should always be based on our specific biological needs so that we select the most suitable ST method to address our research questions. For example, when we want to investigate how the TME regulates treatment effects before and after therapy but lack prior knowledge about which cells or genes are key regulatory elements, we choose Visium HD, which can capture the entire transcriptome. This contrasts with using Xenium, which uses predefined gene panels. However, as imaging-based ST technologies like Xenium continue to expand their gene panel capabilities while offering much larger detection areas than Visium HD, they are becoming more attractive for specific applications. For instance, imaging-based ST platforms like Xenium are a great choice when working with clinical research samples, such as tissue microarrays, where we aim to analyze dozens of samples within a single detection area simultaneously. They not only reduce batch effects but also lower costs, making them ideal for large-scale studies. Do you have any advice or learnings to share with people interested in getting started with spatial? For a researcher who is new to spatial omics, the most important step, as I mentioned in my previous response, is to clearly define your scientific question. Understanding the characteristics of each technical “For instance, imaging-based ST platforms like Xenium are a great choice when working with clinical research samples, such as tissue microarrays, where we aim to analyze dozens of samples within a single detection area simultaneously. They not only reduce batch effects but also lower costs, making them ideal for large-scale studies.” Visium HD spatial profiling identifies the cellular composition in KRAS-mutant colorectal cancer samples treated with total neoadjuvant therapy in combination with bevacizuma. Image credit: Ye lab. 10x Genomics Perspectives from Global Pioneers in Spatial 20 platform is crucial, as the appropriate choices should be based on which one helps you achieve your specific research goals. Additionally, it is essential to adjust the experimental design based on the type of data that the available platforms can provide. It is also essential to familiarize yourself with the unique features of your omics technologies. For instance, if an experiment requires spatial transcriptomics and spatial metabolomics or spatial proteomics from neighboring tissue layers, the approach (e.g., what sample type to use) may need to be adjusted accordingly. For a study involving spatial transcriptomics and spatial metabolomics, it is crucial to use freshly embedded samples. In this case, using fresh samples with the Visium HD [WT Panel] assay would be the optimal choice. However, if the study involves ST and antibody-based spatial proteomics, using formalin-fixed paraffin-embedded samples with the Visium HD [WT Panel] assay would be a better option. Platform flexibility allows researchers to tailor their approach to the specific requirements of their research, ensuring the best possible outcomes. Are there any new directions or capabilities you are particularly excited to see emerge in spatial biology? I would be particularly excited if future technologies enable the integration of spatial multiomics—including spatial epigenomics, transcriptomics, proteomics, and immunomics—on the same slide or neighboring slides. I think the successful commercialization of such multiomic technologies would make them accessible to more laboratories, greatly expanding their use in research. Additionally, I look forward to the broader application of 3D spatial omics, as this would provide even more comprehensive insights into the complex interactions within tissues and their spatial organization and advance our understanding of biology at unprecedented resolutions. 10x Genomics Perspectives from Global Pioneers in Spatial 21 Shaping future therapeutic possibilities: Dr. Freytag on the power of spatial insights in clinical research What is your lab’s research focus? I co-lead a lab with a unique arrangement. There are three of us in the lab: I’m a computational biologist, Dr. Sarah Best is a cancer biologist by training, and Dr. Jim Whittle is a neuro-oncologist. That probably hints at what we do, which is thinking very deeply about brain cancer. In particular, we focus on the development of new Saskia Freytag, PhD Laboratory Head, Brain Cancer Research Lab The Walter and Eliza Hall Institute of Medical Research Dr. Freytag is the head of the Brain Cancer Research Lab at The Walter and Eliza Hall Institute of Medical Research. Alongside research oncologist Dr. Sarah Best and clinician Dr. Jim Whittle, she investigates molecular features of brain cancer that can inform the development of future therapies. 10x Genomics Perspectives from Global Pioneers in Spatial 22 personalized therapies that can fundamentally shift outcomes for individuals with brain cancer. We do this by harnessing Sarah’s and my expertise investigating molecular and cellular features. When we say molecular, we’re extremely interested in the metabolism that defines these tumors, with the goal of translating what we find there into meaningful clinical advances. The work we do is firmly grounded in a patient-centered approach. In practice, this means that we have many programs that go from bench to bedside. We’re trying to develop more accurate diagnostics and also discover new therapeutic strategies for these really aggressive tumors. We then try to push that into our globally unique clinical trials platform that is led by Dr. Whittle and established with help from the Victorian Government. We have pioneered an approach where we’re sampling pre- and post-treatment biopsies. Subjects in these trials undergo a biopsy, receive treatment for a certain duration, and then come back and get a full tumor resection. These precious biopsy samples that we’re collecting for research—these matched samples—really allow us to investigate, with an unprecedented detail, how therapies work on a molecular level. We can also get insights into how they don’t work. That then feeds back into our basic science pipeline where we’re designing and developing new therapies. So it is this iterative loop that we’re trying to do, which hopefully means that we’re finding ways to make care for brain cancer patients much more effective in the future. What types of insights into therapeutic mechanisms have you observed? We’ve only been up and running for around three and a half years as a lab. We haven’t had any compounds or therapies that we’ve developed actually go into clinical trials. However, we have done clinical trials. So we’ve been able to work in parts of the [drug development] pipeline, but have not yet taken a compound all the way. Immunofluorescence staining post-acquisition of Xenium spatial transcriptomics, revealing cell morphology (green = GFAP protein expression). Sample represents a resected tumor from an individual with a low-grade glioma. Image credit: Brain Cancer Research Lab, The Walter and Eliza Hall Institute of Medical Research. 10x Genomics Perspectives from Global Pioneers in Spatial 23 Recently, we’ve been fortunate to finish the first clinical trial that went through this perioperative platform design. It was for an IDH [isocitrate dehydrogenase] mutant inhibitor. Mutations in the IDH enzyme have been linked to many gliomas and alter DNA hydroxymethylation, gene expression patterns, cell differentiation, and the tumor microenvironment. It was really fascinating to use these matched tumor samples to learn how this inhibitor used to treat gliomas works. I think our research could really help shed some light on this. We also believe it will help us understand the mechanisms of some off-target effects that we’ve observed, which seem to be beneficial for some individuals in terms of relieving seizures that are associated with gliomas. Why did you pursue spatial transcriptomics for your research? When we started, we were really aware that metabolomics and metabolite activity in brain cancers is something that has been understudied and could potentially be utilized in [the future] to treat patients. So we’re really interested in establishing that platform in combination with spatial transcriptomics because I think that those things need to go hand in hand. Additionally, while there are great spatial proteomics platforms out there, they are limited in terms of the markers that you can look at. Because we were interested in the composition of our samples and with brain cancer being extremely heterogeneous, with cell populations that are pretty similar to each other, it was obvious to us that a proteomics-limited panel wouldn’t let us see that diversity. Having access to a spatial transcriptomics method, like Xenium panels that look at 500 genes, allows us to be able to delineate all these populations properly. These spatial technologies are essential for our understanding of brain cancer because these tumors are so highly heterogeneous and they infiltrate into the normal brain, which is a tissue that is already poorly understood. The traditional bulk or single cell approaches obviously lose critical spatial context that is an essential component in shaping tumor behavior and this treatment response that we’re investigating in our clinical trials. By preserving this, what we have been finding in our clinical research samples, using the Xenium platform, is that we can now begin to understand how tumor cells interact with the microenvironment and how they respond differently to therapy depending on what microenvironment they are in. Additionally, I think what’s really powerful about this platform is that we can actually overcome some of the drawbacks in these types of clinical trials, where sample variability is something that we’re always concerned with. Whether we sample at the tumor edge or more in the center of the tumor is something we can’t control in a very complex clinical trial design, where subjects undergo a very aggressive sort of surgery. We don’t always know what the surgeon will be able to get. But the spatial context of what tissue samples we do get, actually allows us to interpret similar regions, pre- and post-treatment, and see what has changed in them without being confounded by the cellular composition. Dissecting how gliomas respond to IDH inhibition Although IDH mutations are known to be a key driver of many low-grade gliomas, treatment with targeted IDH inhibitors is relatively new. Using tumor biopsies from the first ever single-arm perioperative trial (NCT05577416) for IDH inhibition, Drummond et al. leveraged Xenium imaging-based analysis to map the TME of matched pre- and post-treatment tumors. This spatial analysis was critical in understanding how treatment changes the TME to alter synaptic signaling, suggesting a potential mechanism by which IDH inhibition reduces seizures. Check out this trailblazing clinical research in Nature Medicine > 10x Genomics Perspectives from Global Pioneers in Spatial 24 In one of your recent papers, you compared several commercially available spatial options and concluded that the Xenium platform was more compatible, more useful. Can you share a bit more about that? We did a lot of comparisons when we first started trying to narrow down the spatial platform we wanted to use. We knew we wanted to be a lab that uses this really heavily, and key considerations for us were what sort of resolution we could get, the gene panel flexibility, the reproducibility of the platform across samples, and, of course, the compatibility with our clinical trial workflow. In the end, we were simply most impressed with the reliability and the data quality we got from Xenium. We’ve now processed over 50 samples in the lab and have had minimal technical failures. I think we’ve had one in all of this time, and that’s probably because the sample quality wasn’t great, so nothing to do with the instrument. The instrument is super reliable. That’s really important when you’re doing a clinical trial and the research sample that you have is really precious. We don’t want to put these samples on an instrument that isn’t reliable and doesn’t give us great data that we want to analyze in the end. These individuals participating in the clinical trial have agreed to undergo two brain surgeries to allow for tissue samples for this research to be collected. We want to be sure that those samples go to maximal effect. Do you have any advice or any top learnings for people that are interested in getting started with spatial transcriptomics? It can be pretty overwhelming to start, and what I would have liked to know is that you make use of every slide in a maximal way by putting multiple tissues on it. I think we’ve now stretched this to using even the tiniest speck of empty space by putting some of our neurosphere or organoid models on there. When you have free space, it’s wasted space, so you might as well utilize the entire imageable area. In terms of what genes to look at, I really love custom panels. I’m a big fan of them because they keep costs down while focusing on the biology that you’re interested in. However, it can be overwhelming to start designing one. But I think once you get your feet wet, you’ll be amazed by how much you find. Finally, I have been a big proponent that cell segmentation isn’t as important as people may believe. Especially if your primary interest is in the tissue composition and cell identities. In this case, I think you can get away with just using the nuclei or the broader watershed segmentation that 10x Genomics pioneered at the start. I say this because cell segmentation is hard, and you’re going to spend a lot of time doing this when simplicity might get you exactly the same information. In fact, we’ve got a publication on bioRxiv at the moment, where we compared different types of cell segmentation. We looked at nuclei segmentation and an actual cell segmentation using a post-staining approach that we did in the lab. This was prior to Xenium making multimodal cell segmentation available. What we found was that the nuclei segmentation and the other approaches were essentially very similar when calling the composition of the sample. The manual cell segmentation staining approach we had developed, however, was taking us hundreds of hours to compute and also quite a significant amount of labor to get done. So, the question is then: is it worth it? There are applications where that absolutely is necessary, but I think people need to really carefully think about if their application is one where it’s truly needed. “The instrument is super reliable. That’s really important when you’re doing a clinical trial and the research sample that you have is really precious.” 10x Genomics Perspectives from Global Pioneers in Spatial 25 What kind of new directions or capabilities are you particularly excited to see emerge in spatial biology? I’m really fascinated with the epigenetic features that we can now look at spatially. I would love for there to be a good platform that allowed spatial analysis of DNA methylation and did it cheaply. I don’t think we’re quite there yet, but, for brain cancer, the epigenome is crucial in defining these tumors. It also probably plays a big role in their plasticity and how they respond to treatment. We also think these processes are driven by the niches these cells live in. Being able to observe that would be really cool. There are so many open questions in brain cancer surrounding the epigenetic markers that have been found to show prognostic value and whether there are areas in the brain tumor where they are really abundant or absent. What do you think your future research direction will be? We have several trials starting this year focused on super aggressive brain cancer, which almost always kills you within one year. In those studies, we’ll be using immunotherapies that are, for the first time, available to patients in Australia. It’ll be really interesting to investigate how the immune system reacts in the tumor itself following these treatments, which is something we haven’t been able to research before. “It can be pretty overwhelming to start, and what I would have liked to know is that you make use of every slide in a maximal way by putting multiple tissues on it.” 10x Genomics Perspectives from Global Pioneers in Spatial 26 Spencer Watson, PhD Senior Research Fellow, Biomedical Data Science Center University of Lausanne, Switzerland Dr. Watson is a Senior Research Fellow at the University of Lausanne’s Biomedical Data Science Center. As a postdoctoral researcher in the laboratory of Dr. Johanna Joyce, he investigated the cellular mechanisms responsible for glioblastoma recurrence in the hopes of finding better treatments. Unveiling tumor secrets: Dr. Watson’s spatial multiomics approach to glioblastoma research What made you decide to use spatial transcriptomics for your research? For me, it was primarily a question-driven choice. I had very specific questions about the post-treatment tumor microenvironment niche and I wasn’t getting answers from disaggregated single cell omics analyses. It was critical that we saw what was happening and where it was happening. 10x Genomics Perspectives from Global Pioneers in Spatial 27 How would you say spatial has impacted your research? It was very transformative for my project. Especially going from looking at disaggregated data and having all these questions about, “Is this actually happening in regions of fibrosis? Is it happening anywhere near the tumor cells?” Being able to look at an image, combine protein expression and gene expression, and everything all in one to say, “Yes, we know what is happening, and we know exactly where it’s happening.” That has allowed us to uncover some unique biology in our latest analysis of the brain tumor microenvironment. Tell us a little bit more about your latest research findings. Our recent Cancer Cell article, “Fibrotic response to anti-CSF-1R therapy potentiates glioblastoma recurrence,” follows up on previous findings where we were using anti-CSF-1R inhibitors to re-educate macrophages in glioblastoma from a pro-tumor phenotype back into an anti-tumor phenotype. It was really successful in massively shrinking these huge tumors in the mouse model. However, we kept seeing rebound tumor recurrence. The super interesting thing we found is that, every time there was a recurrence, it was always associated with a fibrotic glial scar. In cancer research, you don’t get a lot of 100% phenomena, so whenever you see that you definitely need to follow up. Our idea was that maybe these fibrotic scars are causing—or at least potentiating—recurrence, but this is a very complicated environment. It’s very dense. There are a lot of cellular players involved. We needed a lot of different omic modalities. We had single cell RNAseq, spatial proteomics, mass spectrometry proteomics, and spatial transcriptomics. The particularly useful thing for us was integrating all these together to create a multimodal spatial dataset that allowed us to identify that the cellular mitigator of this fibrotic niche was encapsulating tumor cells. Essentially, it was protecting them and driving them to a state of dormancy where they’re safe from other treatments and was protecting them from immune surveillance. With the spatial omic dataset, we were able to identify what the precise cell type was that was causing this and what signals were activating it. That allows us to build a drug treatment strategy to block the cellular response. And our hypothesis was correct. Eliminating fibrosis massively improved the efficacy of the immunotherapy. That’s something that we are very much looking forward to moving forward with in other clinical directions. The future of this project and what we’re doing now is translating this into the patient setting. We found evidence for it in the patient setting in the paper, and now we want to go all-in on patient samples and massively increase the scale because what we’ve seen already is there’s so much heterogeneity there. By going high throughput and high content with that same spatial multiomic depth, we [hope to] be able to identify other combination treatments that can counter the negative side effects of many of the treatments we use in the clinic. What considerations did you take into account when choosing which type of spatial technology you would like to use? The most significant consideration, for me, is that it had to be single cell and, ideally, it had to be nondestructive and image-based so that it could combine with some of our other modalities. “I had very specific questions about the post-treatment tumor microenvironment niche and I wasn’t getting answers from disaggregated single cell omics analyses. It was critical that we saw what was happening and where it was happening.” 10x Genomics Perspectives from Global Pioneers in Spatial 28 Second was how robust the data was. We saw a lot of other approaches that could analyze more genes, but then there were too many questions about the quality of the data you were getting. I think having something that gives fewer genes, but gives you absolute faith and confidence in the data you are getting, is essential. What advice or top learnings would you share with people interested in getting started with spatial? Is there anything you would have liked to know before you started spatial yourself? First, I want to emphasize how important it is to have single cell data to complement the spatial transcriptomic data. Having that combination and synergy gave us the ability to achieve more transcriptional depth and project spatial signatures onto the disaggregated data. That opened up a lot of doors for us. Next, I would say to never do it as part of a revision, mainly because of the amount of interesting questions that you can answer once you have it. Spatial is something to do at the start of a project rather than at the end; otherwise, you’ll add several years to your analysis chasing down all the interesting leads. Also, make sure that you actually have a question that is applicable to spatial. I think a lot of people feel that they have to do it for grant purposes or something similar. But, if they don’t have a defined question that spatial resolution will answer, they can spend a lot of unnecessary analysis time trying to make use of the data. Are there any new directions or capabilities you’re particularly excited to see emerge in spatial biology? There is a lot of focus on spatial proteomics. It was the method of the year; previously, it was spatial transcriptomics. But I feel like it’s all going to be about multiomics, about how many data types can you integrate, so that we can ask these computational systems biology–level questions of emergent features that you would never get from even one of these powerful approaches. I think increasingly you’re going to see protein with transcriptomics and metabolomics. Spatial ATAC-seq will probably come along as well. The idea of just doing one of these may seem like an anachronism in five years. Reimaging our understanding of glioblastoma dynamics More than 90% of patients with glioblastoma, the most common and aggressive brain cancer, have their tumors recur even after treatment. Watson et al. integrated Xenium single cell spatial imaging insights with proteomics and traditional single cell RNA-sequencing to define the association between fibrotic scars and glioblastoma recurrence. Critically, these high-resolution multiomic insights enabled the identification of a multidrug treatment regime to limit scar formation and, therefore, boost treatment efficacy, which showed promising preclinical results. See the stunning data that earned a cover spot in Cancer Cell > “Being able to look at an image, combine protein expression and gene expression, and everything all in one to say, “Yes, we know what is happening, and we know exactly where it’s happening.” That has allowed us to uncover some unique biology in our latest analysis of the brain tumor microenvironment.” 10x Genomics Perspectives from Global Pioneers in Spatial 29 What do you feel is the biggest benefit of doing all these modalities simultaneously? Especially in the cancer field, we are working with very precious samples, and we really need to get the most out of them. This sample was a major contribution from a patient. We should make sure that we are asking every possible question and getting every possible thing that we can get out of it, not just for us but also so the people who come after us can continue to use this data. By making these massively deep datasets and making them publicly available, it has this snowball effect where we asked our question of it, but there are many other researchers who can now ask their questions of the same data. Is there anything else you would like to highlight? I think the most exciting thing now is scale. Cancer is so heterogeneous that doing 3 or 4, even 10 or 20, samples barely scratches the surface of the variability. When we scale this up to hundreds and beyond, that’s when the AI revolution becomes really exciting, because then we can train AI models that saturate all the noise in cancer and overcome this burden of heterogeneity. What these AI models can find that we’ve never appreciated, that’s going to be the really exciting stuff in the next 5 to 10 years. 10x Genomics Perspectives from Global Pioneers in Spatial 30 Make your spatial goals a reality As the incredible researchers featured in this eBook have shown, spatial transcriptomics enables powerful insights and opens up new areas of investigations. If you are ready to journey into the world of spatial transcriptomics and gain new levels of understanding in your research, we’re here to help turn your spatial goals into real breakthroughs. The 10x Genomics Xenium Spatial and Visium Spatial platforms have been designed to let researchers like you explore tissue landscapes with exceptional precision, flexibility, and reliability. Get to know a little more about each platform below, then follow the QR codes to continue your journey. Visium platform A whole transcriptome spatial gene expression platform, Visum expands single cell–scale spatial discovery with the broadest view of gene expression. Why choose Visium? Whole transcriptome analysis with single cell–scale resolution Broad discovery power with a 3’ poly(A)-based assay & NGS read-out Working with archival or pre-sectioned slides Xenium platform A best-in-class ultra-high plex in situ technology, Xenium delivers robust single cell spatial discovery with unmatched customization. Why choose Xenium? Subcellular resolution & higher per gene sensitivity Custom panels for targets & species of interest Large imageable area provides flexibility to run multiple tissue sections per slide Scan the QR code to learn more about the Xenium platform. Scan the QR code to learn more about the Visium platform. 10x Genomics Perspectives from Global Pioneers in Spatial 31 Selected publications and resources Dr. Miguel Reina-Campos Reina-Campos M, et al. Metabolic programs of T cell tissue residency empower tumour immunity. Nature 621: 179–187 (2023). doi: 10.1038/s41586-023-06483-w Reina-Campos M, et al. ImmGenMaps, an open-source cartography of the immune system. Nat Immunol 26: 637–638 (2025). doi: 10.1038/s41590-025-02119-5 To access ImmGenMaps datasets, visit https://www.immgen.org/ImmGenMaps/ Dr. Amanda Poholek Spatial cytokine microniches direct T helper cell pathways that drive allergic asthma. Nat Immunol 25: 1999–2000 (2024). doi: 10.1038/s41590-024-01987-7 Dr. Nicholas Banovich Amancherla K, et al. Dynamic responses to rejection in the transplanted human heart revealed through spatial transcriptomics. bioRxiv (2025). Mallapragada S, et al. A spatial transcriptomic atlas of acute neonatal lung injury across development and disease severity. bioRxiv (2025). Dr. Youqiong Ye Du Y, et al. Integration of pan-cancer single-cell and spatial transcriptomics reveals stromal cell features and therapeutic targets in tumor microenvironment. Cancer Res 84: 192–210 (2024). doi: 10.1158/0008-5472.CAN-23-1418 To access the SpatialTME database, visit http://www.spatialtme.yelab.site/ Xun, Z., Ding, X., Zhang, Y. et al. Reconstruction of the tumor spatial microenvironment along the malignantboundary- nonmalignant axis. Nat Commun 14: 933 (2023). doi: 10.1038/s41467-023-36560-7 Learn more about Cottrazm at https://github.com/Yelab2020/Cottrazm Dr. Saskia Freytag Kriel J, et al. An integrative spatial multi-omic workflow for unified analysis of tumor tissue. bioRxiv (2024). Causer A, et al. SpaMTP: Integrative statistical analysis and visualisation of spatial metabolomics and transcriptomics data. bioRxiv (2024). Dr. Spencer Watson Watson SS, et al. Microenvironmental reorganization in brain tumors following radiotherapy and recurrence revealed by hyperplexed immunofluorescence imaging. Nat Commun 15: 3226 (2024). doi: 10.1038/s41467-024-47185-9 Contact us 10xgenomics.com | info@10xgenomics.com © 2025 10x Genomics, Inc. FOR RESEARCH USE ONLY. NOT FOR USE IN DIAGNOSTIC PROCEDURES. LIT000229 - Rev B - Spatial Leadership eBook References 1. Coons AH, Creech HJ & Jones RN. Immunological properties of an antibody containing a fluorescent group. Proc Soc Exp Biol Med 47: 200–202 (1941). doi: 10.3181/00379727-47-13084P 2. Pardue ML & Gall JG. Molecular hybridization of radioactive DNA to the DNA of cytological preparations. Proc Natl Acad Sci USA 64: 600–604 (1969). doi: 10.1073/pnas.64.2.600. PMID: 5261036 3. Singer RH & Ward DC. Actin gene expression visualized in chicken muscle tissue culture by using in situ hybridization with a biotinated nucleotide analog. Proc Natl Acad Sci USA 79: 7331–7335 (1982). doi: 10.1073/pnas.79.23.7331 4. Williams CG, et al. An introduction to spatial transcriptomics for biomedical research. Genome Med 14: 68 (2022). doi: 10.1186/s13073-022-01075-1 5. Asp M, Bergenstråhle J & Lundeberg J. Spatially resolved transcriptomes—next generation tools for tissue exploration. Bioessays 42: e1900221 (2020). doi: 10.1002/bies.201900221 6. Moses L & Pachter L. Museum of spatial transcriptomics. Nat Methods 19: 534–546 (2022). doi: 10.1038/s41592-022-01409-2 7. Method of the Year 2020: Spatially resolved transcriptomics. Nat Methods 18: 1 (2021). doi: 10.1038/s41592-020-01042-x