Interview With Dr Doyle - Father of Spatial Biology.
- arbrady
- Feb 17, 2024
- 9 min read
Updated: Feb 28, 2024

The inability to analyze the landscape of a tumor or even create a detailed characterization of specified tumor landscape was a huge block in the development of a cure for cancer. To combat this hurtle, Researchers have begun looking into applications of Spatial Biology, an area of biology that focuses on the 3-D mapping of cell morphology. Before the 1990s it was a completely different story. In order to analyze an RNA sequence, Biologists had to break down tissue to the cell in order to get data. Doing so is not just time consuming, but also causes the biologist to lose any spatial context of the cell and how it interacts with the cells around it. On the other hand, one could examine the morphological features of an organism, but the structural context would be lost. These two process remained exclusive with no way to bridge the gap. That is, until the invention of Spatial Biology. Now, a Biologist is able to simultaneously observe a tissue and the cells that make it up. Spatial Biology is at the forefront of mapping the landscape of disease, the study of heterogeneity, and cancer precision and detection; with applications that are expanding as technological advancements in the field continue. But how did we get here and what did it take to create this new field of Biology?
Allison
Dr. Doyle, they call you the father of spatial biology. Why is that?
Mike Doyle
Well, back in the mid-1990s, I started a project called the Visible Embryo Project and I was on the faculty at University of Illinois at Chicago. One of the outcomes of that project was by in the later that decade.
So in 1998-99, our team got together and designed the first system for what today is called Spatial Transcriptomics, which is the foundation of what is known as spatial biology today. We filed the first patent for that technology in 2000 and that issued officially as a patent in 2009, but the technology really didn't start taking off until the rest of the world discovered it in the following decade around 2016.

Allison
So can you explain what Spatial Biology is?
Mike Doyle
Yeah. It's an approach and set of methods to both capture the fine structural detail relating to tissue morphology and capture the signals necessary to map biological function with fine localization detail on that tissue morphology. So you can not only see what's happening, but you can see where it's happening at even the cellular level.
Allison
Alright. So what was keeping that from happening beforehand?
Mike Doyle
Well, there were a number of technologies that had just been developed. One of the key technologies was the first gene chips that would allow you to study a sample and look to see what the activity level was for all known genes, for example, the expression activity. Those systems were designed for taking tissue and they would take a small block of tissue and you'd liquefy it and then process that liquefied sample and capture the signal activity from that, but in doing that you'd lose the biological structure. In order to do this kind of work, you'd need to have a tremendous amount of computational power to capture the structure and map the signal onto the structure. Then provide an interface so you could navigate through that data to see what's happening and where it was happening.
This was in the mid 1990s when, you know, the computers that were available at a personal cam computer level were less powerful than your smartwatch today, probably by a factor of a thousand. And the larger computers that you would need to even begin to work with these large datasets that would be produced from this would be in that era, they were supercomputers. So we had to create new technologies to make this possible. The technologies didn't exist at the time, we had to invent them.
Allison
I read up that one of those technologies that was invented for this project, it is known today as the "cloud" system. How was that implemented in your research at the time?
Mike Doyle
Well, part of the idea was to create an online resource. At the time, the very first generally available web browsers were starting to emerge. There was a famous browser from the University of Illinois called "NCSA Mosaic", and the web was really simple at the time. You could display documents with static images in them and you could link from page to page and that was about it. I saw that as an interesting way to make a universal interface that could be used to interactively, you know, access really powerful remote computational resources if we developed a new kind of web browser and a new kind of web infrastructure. So we wound up coming up with a kind of browser where you could run a small program.
So consider how a program like this would be distributed, you have pieces of it running all over the place; running locally, running on the network, running on remote computers with all these pieces of one big program. Well, the pieces running on the on your local computer could be running in the browser and it could display data in a web page, but running real time data, was not something that had been previously created. That could communicate with other pieces running on remote, a whole bunch of remote computers that would calculate based on these enormous datasets, would calculate views of that data in response to your interactive commands. That's what is today called the cloud., and that's how the cloud works.
You are, you know, issuing commands or controlling actions on the browser, but those are actually just being sent to a cloud of a huge number of remote computers that are calculating the result and streaming that back to you then displaying them in the web page. That didn't exist at the time, and we had to create the first version of that.
Allison
And how did you guys implement that into your research?
Mike Doyle
Well, that became the basis for the, project that was called the Visible Embryo Project. So in the early 2000s, we put together a project with a number of institutions across the country to build a system that could allow people to interactively navigate that kind of data and annotate it and work with it.
Allison
So while talking about this, you've talked a lot about the Visible Embryo Project. What is that and what was your involvement in the project?
Mike Doyle
Well, there was in the very early 1990s, the National Library of Medicine had created a project called the Visible Human Project. The idea was to take a cadaver and slice it up then digitize it at such high resolution that you could reconstruct the anatomy in incredible detail. I was appointed to the oversight committee for that project and it became clear that it would take years before that data was available. I started looking for other sources of human anatomy data that could be digitized and reconstructed because I wanted to work with that kind of data to see what we could do with it.
I came across a collection of embryology specimens from the National Museum of Health and Medicine in Washington, DC called the Carnegie Collection of Human Embryology. It was essentially 650 little Visible Human Projects. There were these embryos that were sectioned out and they where mounted on slides.
Her name was doctor Adrienne Noe, she later became the director of the museum. At that time, she was running that specific project, and she's still the director today. While I was working there, we got together and started talking about what we might be able to do when the idea came up to create an online resource for human developmental anatomy. From the very beginning, one of the ideas was we could use that to create these canonical models of human development that could be used to do things like map the results of expression data based on what, at that time, had been recently released in terms of the results of the of the Human Genome Project.
So the thought was this knowledge base could be a way to pull all that kind of data together and start mapping out what later became called spatial biology information.
Allison
So what did this early technology look like?
Mike Doyle
The idea was basically that you take a biological specimen, you image that specimen at very high resolution, and then you come up with a way to subsample that specimen. Basically, you're rasterizing the tissue in some way so that you can analyze each of those subsamples independently and ideally and be able to search for the activity of all known genes in each of those locations. So each of the specimens would be isolated and barcoded, analyzed, gene expression data would be stored, and then remapped back onto the original image data. So each three-dimensional pixel, which or otherwise, it's called a voxel. Each voxel of information in the original image would then have within it the results of the presence of, or lack of presence of all known, you know, mRNA activity. And with and so you're it's basically mapping what's today known as the transcriptome, but spatially in the tissue. And so the final steps then are to, you know, consolidate that into a single database that can then be visualized and navigated and studied and analyzed.
(Error with the recording system occurred)
Allison
We were just talking about the early technology of it.
Mike Doyle
Right. And you're asking what it looked like. So I know I was talking about this method. There were these various steps.
So you basically isolate you image the tissue, subsample the tissue, analyze each of the subsamples, and in that case, each of those would be run up on a PCR chip and then analyzed using gene expression analysis. At the time, one of the methods we described would use gene chips to analyze the expression activity. More recent modifications of this use other methods where you take an oligonucleotide sequence that is bound on a grid to a slide and has a tail that will then capture whatever mRNA that matches. The section's laid on top of this grid of all these oligos and the mRNA is captured then the tissue can be removed and it can be processed for the signal, which is the map back into space. It's essentially the same process just using a few different steps.
Allison
And what are the benefits of using this new method versus the gene chips?
Mike Doyle
It's much cheaper. Much I mean, the earlier methods would require robotically separating and isolating the individual samples. There'd be advantages in terms of higher resolution, but the newer methods are much faster and much cheaper, much more practical. The paper that was published in Science by the Sweden group in 2016 described that kind of method, which is part of the reason why it, I think, took off because it was a practical method that could be reproduced by others.
Allison
Developing this new method of biological analytics must have come with its ups and downs. What are some of the difficulties that you and your colleagues faced while working on this?
Mike Doyle
Well, the biggest problem is the massive amount of data. If you think about this kind of method, you could capture petabytes worth of data from a single specimen, from a single slide if you capture at the highest resolution that's available today. We, even today, don't have all the tools necessary to really get the most out of that kind of data. So a lot of the current activities are focused on new computational methods for doing spatial analysis, data visualization and higher order statistics looking at finding patterns in that data. By far, the biggest problem, I think, today is the massive amount of data that's generated by the process.
Allison
How did your team confront this problem when you guys were just getting started?
Mike Doyle
Oh, we had to create the cloud. The cloud is that's where that was the motivation that led to the creation of the cloud technology.
Allison
Now looking at the modern day spatial biology field, there's been a huge incline of papers being made about this subject. In fact, spatially resolved transcripts were labelled the "Method of the Year" by Nature in 2020. So, what do you see as the future for spatial biology?
Mike Doyle
I think you're just gonna see it appearing more and more often. It's the new tool and there are breakthroughs that are coming from the new knowledge that's being generated. I'm particularly interested in looking now at how to pull together all the data that's being generated and see if we can create common resource repositories that people can use to mine new information from these massive amounts of data. So there are experiments being structured to go after particular questions hypothesis driven experiments, but the data that's being generated is so rich that I think there's a lot of opportunity for new discoveries that'll come through just exploration of those data sets. That's not particularly hypothesis driven. It's gonna result from people studying these data and finding patterns that they weren't looking for initially.
Allison
So this field looks like it's not slowing down anytime soon.
Mike Doyle
No. I think it's only gonna accelerate.
Allison
Alright. Well, thank you so much for giving the insight to this, and thank you so much for doing this interview with me.
Mike Doyle
Oh, it's my pleasure.
Resources
Deoxyribonucleic acid (DNA) fact sheet. Genome.gov. (n.d.). https://www.genome.gov/about-genomics/fact-sheets/Deoxyribonucleic-Acid-Fact-Sheet
Infante, D. (2023, October 4). Clinical applications of spatial biology in disease diagnosis and treatment. News. https://www.news-medical.net/health/Clinical-Applications-of-Spatial-Biology-in-Disease-Diagnosis-and-Treatment.aspx
Marx, V. (2021, January 6). Method of the year: Spatially resolved transcriptomics. Nature News. https://www.nature.com/articles/s41592-020-01033-y
Schneider, E. (2023, May 26). The Birth of Spatial Genomics. NanoString. https://nanostring.com/blog/the-birth-of-spatial-genomics/
Nice, Allison!