This is a machine transcription and therefore it may contain inaccuracies, errors, or mispronunciations. Notice an error you think needs changing? Please contact the Bitesize Bio team using this form: https://bit.ly/bsbtranscriptions

Intro/Outro (00:00:02):
Welcome to The Microscopists, a Bitesize Bio podcast hosted by Peter O'Toole, sponsored by Zeiss Microscopy. Today on The Microscopists.

Peter O'Toole (00:00:18):
Today on The Microscopists, I'm joined by Mark Ellisman of UCSD and we discuss his early ambitions.

Mark Ellisman (00:00:26):
So you might say I, in my early years, the most interesting thing to me was to make a chop shop for bicycles,

Peter O'Toole (00:00:34):
The importance of generosity in scientific collaborator

Mark Ellisman (00:00:39):
That no single investigator should receive so much money to have all of that stuff to themselves.

Peter O'Toole (00:00:45):
The pros and cons of deep learning,

Mark Ellisman (00:00:49):
The deep learning is sort off easier to to test with regard to accuracy on test data sets. The problem is the expense of competing

Peter O'Toole (00:01:00):
And his advice to early career scientists

Mark Ellisman (00:01:03):
For the, those who, who want to avoid feeling like they're in competition all the time, trying to ask questions that are legitimate, but two steps beyond everybody around you

Peter O'Toole (00:01:18):
In this episode of The Microscopists. Hi, and welcome to this episode of The Microscopists on Peter O'Toole from the University of York. And today I'm joined by Mark Ellisman, the National Center of Microscopy and Imaging Research at UC San Diego. Mark, how are you today?

Mark Ellisman (00:01:42):
I'm well, I'm well enjoying nice weather here in California. I hope you're doing the same, Peter.

Peter O'Toole (00:01:50):
It's okay today. And that's very, for British, that's really positive if I say it's okay. I'll. I haven't asked this as a first question. What was the first microscope you ever used? Can you remember which microscope it was? What type?

Mark Ellisman (00:02:13):
Since I started out doing electrophysiology, I think the first microscope I probably used was a compound microscope, Right Field light microscope, the type who knows, but it probably was American optical or something. Very simple.

Peter O'Toole (00:02:32):
And I remember my first experience. Did you actually like the microscope that you used?

Mark Ellisman (00:02:42):
Peter I don't want to extend this, but like is a hard question in the sense of it was extremely simple. There was not much more than brass and glass, a light source. If I remember correctly with a cord that needed repair I was not a microscopist when I started out being curious about biology. So that was not the centerpiece of my activity. The microscope was more interesting to me.

Peter O'Toole (00:03:25):
I don't think that's uncommon. I think so many people who recognize from microscopy or use or development of microscopy never started with an interest in the microscope. It was they, they found their way into it. So what was your, what did you want to be as a child when you grew up?

Mark Ellisman (00:03:46):
Well I come from a family with a father with a very strong engineering background. He was a aerospace engineer, a nuclear engineer. And so from the time I was very small, I fiddled with my father in the workshop and he taught me a lot of, I would just say how to make things and some of the principles of materials and physics. And so I spent time making bicycles or making go-karts eventually making vehicles of different sorts. Trying to, how would you say manufacturer with his tutelage, those piece parts that made them unique? So you might say I in my early years, the most interesting thing to me was to make a chop shop for bicycles.

Peter O'Toole (00:04:51):
Okay. So did you see your, when you were a child, do you think that's what I'd like to do? And I, when I grow up and

Mark Ellisman (00:04:58):
No, I just figured that that was something I enjoyed the small accomplishments of making something on the weekend or having little construction projects. And I think I realized that that gave me joy, the kind of internal accomplishments of making something. And when I expanded I would just say my thinking to be curious about things like the brain. I decided that I would use what seemed to be a talent for using my hands and my head to make stuff, to try and have at least a little piece of what I did as a scientist involved making things, because that would give me kind of short term sense, you know, sense of accomplishment in the short term. Whereas the bigger questions were quite to get answers to big questions was a lengthy business with lots of Hills and valleys on the way. And I thought I needed a little bit of a day to day sense of accomplishment, and I thought that would come from making the parts to enable the science. I think.

Peter O'Toole (00:06:17):
So what got you into the science to start with? Cause you obviously your degree was in science, then you went on to your PhD, I think at Colorado. What trick?

Mark Ellisman (00:06:31):
I, I did my completed my undergraduate work at Berkeley in the late sixties, which was a bit of a heady time with you know, protests and everything you heard about the sixties. I know you're probably just a little bit younger than me are true, especially if you happen to live in the San Francisco bay area at that time. So I would say that I went to college wondering how the brain works, being impressed by, you know, discoveries that I listened to about on the radio, like, you know, Watson and Crick's breakthrough, you know, which was told very nicely by expose. They have the double helix or whatever, which was reading at the time. I guess I was in junior high school and I decided that the brain would provide a sandbox if you like for exploration for a lifetime or more. And that there would be discoveries if one could put together strategies that would help me to understand, you know, with my kind of sixties curiosity, what was actually going on in my head, why I was me, why I reacted in different ways. What was perception? You know, we were all curious about what an individual's perception was and how fixated it was by the chemistry at the moment.

Peter O'Toole (00:08:18):
I, I just love the thought that, so obviously you went into neuro science and bioengineering and part of your drive is when you're actually developing the tools, you developing the tools, you see a result, you get excited and then you want to see what's going on in your brain. What's causing you to get excited, which drives your next project. Obviously you can't do your own brain. I guess that'd be, I guess that'd be a no brainer at that point. If

Mark Ellisman (00:08:45):
There are a few who would like to donate their brain to get their wiring in the interest of their own immortality, I have had those conversations. The I'm not one who thinks that's the way the immortality, I think one, one needs out of this at some point. So in mortality is not exactly my goal. The I think that the stories, I don't want to make a long answer because we have a limited amount of time, but it's kind of interesting. When I was at Berkeley I completed my requirements for my undergraduate degree, a little early and I was given the opportunity to do some what were called honors activities, which are graduate courses. And that I had been working on animal behavior in one of the labs, looking at how different species of animals could respond with different levels of complexity in their response to something novel in their environment. And then you asked me what was my first microscope? Well, the first microscope was what I had to use to analyze the circuits at a very primitive level in the brains of animals that could manipulate things greatly that were novel in their environment versus ones that were more sloth like in having almost no interest in anything that was novel, unless they could eat it. And so that kind of led me to think that there were other ways besides just cutting up the D and looking at the dead fixed anatomy to monitor. And I took a course from actually Horace Barlow. Who's a famous in the UK. I think Horace is probably still among us, but rather old where there were just six of us. And we learned the classic physiology from cats to rodents or whatever Hubel and Wiesel type physiology.

Mark Ellisman (00:11:03):
So a lot with primitive electrodes, remember this is 1969 or something like that. And I was impressed. And then I actually sought to go to graduate school, to work as a physiologist with electrodes and look at subtle aspects of the nervous system. So what I asked in my first graduate in my first PhD program, I had two PhD programs in Colorado that I joined actually the first one was to ask if you look at a neuron that fires very reliably, when it's asked to like a motor neuron to move your leg muscles, how does it differ in its channel behavior like channel behavior asan channel behavior from a neuron that actually changes its probability of responding if it's been activated previously. So this was at that time kind of one of those holy grail questions about, you know, where does plasticity reside there's interest in, you know, whether it's at a synapse that chemical places where one cells fits neuro-transmitter another with a delay of a couple of milliseconds, the next stage of how neurons work is that activity with the chemistry of the synapse results in a change and the potential across the membrane. And then it fires off an action potential in the signal goes a very long distance, but that's not an highly reliable process. So I was really curious about why some neurons right next to other neurons hold reliability and others hold plasticity, or, you know, the change. And so I worked on that at the wrote software programs used micro electrodes and pretty much outlined the complete thesis on that topic as a intracellular physiologist in the very early days of sharp electrodes. And I decided that looking at the abstraction that was oscilloscopes or reading out graphs and charts, as you would as sort of a physical sciences oriented person using analytics aided by computers was one or two levels of abstraction beyond where I thought I would benefit the most. And there was a summer course on for an electron microscopy by the department next door. I was already pretty far along and finishing my PhD. So I asked my advisor if I could go take the course because I wanted to look at that part of the neuron and see if I could see anything there that would give me a, a charge mentally as to what I might ask. That's different from what we've been asking with the abstraction of a microelectrode probably the problem arose. The course had a fee. My advisor had no money. So being I guess a bit of a high energy kind of guy at that time, I've calmed down a lot. I approached the course coordinator, the chairman of the department, who I didn't know was a famous person, just knew he was the gatekeeper for the course. His name was Keith Porter. And I I went and had a meeting with him to ask if I could get into the course for free, you know, cause I couldn't pay the fee unless I paid it out of my pocket. And that meeting was rather remarkable because I told Porter what I wanted to do. And it turned out that we can serve each other's purposes. He needed someone who was energetic and curious to try and use a microscope that he had just procured. That was 1 million volts. So three stories tall and he'd promised someone who'd given him a a fellowship to award to somebody, a guy by the name of Ernest Fuller, Ernie Fuller, who was with Porter at the Rockefeller Fuller made the first grid when Porter developed the microtome. So he didn't know who to give the fuller fellowship to. And the fuller fellowship was for neuroscience and he hadn't had a neuroscientist come. So he said, well, how are you being supported young Ellisman, would you like a fellowship? And I'd been doing teaching assistant work at that point for three years, which was a bit of a grind. So I said, sure. And that was exciting. Of course, what Porter had in mind was that I would leave the department, I was in and come and do a PhD with him. So he was like, you know, yeah, with the bait out. And it had hooked me, which was good. It turned out that the million volt microscope was not very practical. It couldn't figure out how to stain immediately. And it didn't get me closer to my channels. But down the corridor from the Porter, Porter lab was a guy by the name of Andrew Staehelin, who was one of the early people doing threes fracture, which had the promise of revealing a membrane proteins. I mean, it had with the old membrane proteins and channels. And since I was a hot to look at the voltage dependent sodium channel, which we knew was a channel that nobody purified it or seen it, I thought I could least hunt for the sites where we knew it was concentrated by electrophysiology. So while I was waiting for some help to figure out how to use a high voltage microscope and get contrast, because 1 million volt electrons really don't leave much trail to give you scattering contrast. I went and learned freeze fracture and then did freeze fracture on systems that exhibit plasticity. So eventually we figured out how to use the million volts. And so that's another story we can get to.

Peter O'Toole (00:17:49):
So You're very fortunate. Well, I guess very fortunate that that opportunity arose. And then if you fast forward to your recent work, it's not just the electron microscope. It's the combination of all, of, most of my microscopists, like microscopy, electro microscopy, x-ray microscopy really to study not the same question, but to just dive deeper and understand. And this is still sorta of the same question though, isn't it? It's still trying to understand the basics of the brain, a neuronal system Def.

Mark Ellisman (00:18:27):
Yeah. Peter, I don't want to be to corrected, but the brain has been my passion and the diversity of microscopies that I've managed to drag into at least have a available to, to look sometimes at the same question is because most questions are, multi-scale not only in dimensional terms, but in time. And so if you're looking at a channel with a microelectrode and the abstract signal that you see, you're looking in time domains of milliseconds or fractions thereof, because that's the cadence of their behavior or their dynamism. Yeah. More aggregate properties that would happen with, let's say a thousand channels or receptors or something in a small patch of membrane, let's say a micron, which is easily a micron domain is easily seeing the light microscope, but channels that are 10 nanometers in diameter are not right directly visible with the light microscope. So you need to merge all of these things to try and get some time domain. So grounding and each one of the microscopies that we brought in to, you know, have available at the national center, are we linked to where we can't, you know, field, let's say a multi, I have mass spec we haven't harvested enough money to have one of those of our own we use in combination on the same problem. But the, the key to, I think the success of the center has been really two fold. One is articulating a very simple kind of mantra. And that's that if you combine frontier activities in chemistry, let's say chemistry of labeling how you, how you mark something for different modalities, light or electronic, microscopy with engineering or instrument development. So again, going back to what I said earlier about engineering gratified for making a better bicycle wheel or, or whatever, if I could spend a little bit of time satisfying my amateur engineering niche, right? Learn enough physics to know something about electron optics and camera technology. And then we made a lot of cameras, including the direct detector, you know, at the center. So engineering, so chemistry, engineering. And then the part that I also learned from just using early computers down all the way back to punch cards at Berkeley was that you need to put some math and analytics behind it because you have to turn, what's otherwise likely to be an observational science into something where you can put numbers on top of the data and potentially pitch to people that are braver than me who do simulations to determine from simulations or predictions, which direction to take the next experimental science. So that's kind of been the, the, how would you say that? The way I built the legs on the stool that had held up the center combined with the realization that no single investigator should receive so much money to have all of that stuff to themselves. And so the only way that I could build a center with, you know, $70 million worth of current hardware from time to time was to make sure that we were generous in collaborating with a worldwide community to how would you say justify very expensive tools, keeping them current? That's, that's been our mantra, even though there've been some themes of science,

Peter O'Toole (00:23:09):
I was going to say, you talked about making engineering, but you've also been involved in the manufacturer, the tools you have to think mini solve some of the other bits. And part of those are through collaborations. You've had to be there to, to be at the forefront of a lot of the field, a lot of the time. And that's quite a talent, actually, you, you, you know, you said you've got $70 million worth of equipment, but that doesn't come by luck that comes by, you know, having those collaborations, making it worthwhile, given the impact back to science as well. That how, how have you found that as a challenge that's quite challenging to keep at the forefront for so long. Now you must have a, a great team that's been under you to, to help support that. Cause you, as a lead person, you then have to have people to, to enact and actually become, you know, to really enable so much. How, how have you balanced that?

Mark Ellisman (00:24:13):
Well, the first thing is to realize that you're only as good as the smarter people around you and that it's not just the politics of how you pitched to raise money where, you know, you don't go in with a crown with a bunch of you know, be trucks you go in, you know, with an enterprise or crown that has chimed jewels, people that are highly regarded and viewed as reliable. Now, I've been extremely lucky in that, even though I'm probably not in many of these collaborations, the, how would you say the, the smartest one in the room by a long shot that I've been able to be friends and collaborate with people that are absolutely brilliant. Roger Tsein being one of them, we hit it off. When Roger came to UCSD both of us liking to make things, Roger being a brilliant chemist and having you know, maybe we taught each other a few things. I learned a lot more from Roger than I think he learned from me. But one part of it was me influencing Roger about the value of going from a light, kind of the miracle of light, some in phonics, say to how we would turn the signal that you get from some dynamic probe. I mean his most important work for calcium indicators. He got the Nobel prize for the fluorescent proteins, but all of us were going, here's another example where the most important work was left on the table with regard to the Nobel committee, but how we started a project to try and figure out how to make a probes that would show when one of his calcium indicators, one of the non-genetic ones before JFP was with calcium versus when it was without calcium. My lab, you know, and John Singer, who was a collaborator in early days of the lab, he helped me start Nick Mer singer of the fluid mosiac membrane legacy. We were doing a lot with antibodies and Toki Hossu was part of the team. He was the one who developed super S embedded pro sectioning and immuno labeling. So Roger had the idea that we would make monoclonal antibodies together that would recognize one of the calcium indicators with calcium, without calcium differentiate. And so we could follow up in EM places that had calcium transients would then be marked at higher resolution.

Mark Ellisman (00:27:26):
I was skeptical. I'm not afraid to, in the presence of greatness to say, well, maybe not tell me, tell me I'm wrong. I said, Roger, you know, if we do the antibody stuff, it's not going to be as pretty as we'd like, because you have to permeabilized to get these immunoglobulins in, you know, there and the animator complexes. And so that pristine ultra structure won't be there. We might be better than light microscopy, but you know, it might be ugly. I said, why don't we think about photo oxidation, let's use, you know, one of your molecules or some derivative thereof to generate reactive oxygen. And I think that reactive oxygen we can control the chemistry. You tell me are the expert Roger. So that the paint that we put on the enticement, which is electron dance by, you know, one or another trick we figured out would be, you know, like a five nanometers shell kind of like negative standing insight too. I said, that might work better because then all the reactants are small molecules. Roger, initially, you know like any brilliant person liked his own idea first. Right? So he, he said, well, that could be really muddy. Right. And I said, well, Roger, let me try it. Okay, help me here. I said, you know, I worked a lot with the estro coleen receptor. I had,uworked with purified receptor with Jon Lindstrom and Mike Raftery and we'd characterize the receptor and negative stain and reconstituted it. This is stuff that I did when I arrived at UCSD and I think 1978, 79, I said, why don't we just use [Inaudible] which binds irreversibly to two of the sub units and the nicotinic receptor. And let's put a fluorifor on it, you know, just by, you know, direct conjugation that will generate reactive oxygen. And he, he said, okay, so just use Tetra Bromo fluorescing. And I said, oh, okay. Well, I know about fluorescene what's Tetra Bromo. And he laughed at me like I was his dumber younger brother is, you know, he said, it's eosin. Right? And I said, oh, eosin. I know what it is. So we made those conjugates or I had molecular probes to it. And then what I was able to show at the neuromuscular junction, which I had studied well as a graduate student, was that I could make the estro coleen receptors stand out as individual molecules in a thin section. So by painting for, I had done, when I'd suggested to Roger, we might be able to, to was doing encasement staining of the molecule and,

Peter O'Toole (00:30:35):
And what was missing, what was his take on it when he saw it, this muddy image that he was expecting,

Mark Ellisman (00:30:40):
It was okay. We'll make probes that are reactive oxygen generators. So we, we shifted from the antibodies cell and to reactive oxygen generators at first a series that were, you know, related to you know, no complex molecular biology. And then once the multiple colors, the GFP came around, we shifted to look for other expressible molecules, like mini [inaudible], ultimately, which was based on something from a phototrop and to and, and so we're kind of off to the races. It's still more complicated to do those things than it is to use an enzyme like the apex that we co-develop the Dallas ting, but the apex generates a formulating reaction. And we'll be more muddy. The photo oxidation is still probably the best in terms of a high fidelity localized stain. It's just that it's harder to disseminate because you have to train people to do it. Right.

Peter O'Toole (00:31:59):
So I'm going to switch track a second. You have, if anyone listening, you should actually just tune in just to see Mark's background, because I could have that picture on my wall as a big picture. I would, it is. And just, just the colors that are chosen are beautiful and you send me some other pictures as well. So again, these are all confocal images, right? We're not mistaken, but then you also sent me some of the volume, EM

Mark Ellisman (00:32:34):
Yeah, yeah. This is from what we call the Denka tone. The one we're behind now, that image is of the cerebellum and that's the largest cell body in the brain called the Purkinje cell. And in the background and the kind of binarized gray scale is the rest of the volume imaging data, but just sitting on top of it that, I guess it's a a little color deficient, it's a orangy red hairball. Those are deep learning based segmentation of all the mitochondria. And that's from a project fellow in the laboratory. Matthias Haberl, who's now in Berlin. Did we developed a lot of deep learning algorithms to map sub-cellular organelles going back now, 10 years, something Janelia has taken up in high energy lately and we're happy to use their tools our tools that are publicly available. The one on the right behind me, sorry, is like the one that you had up behind you which is a confocal image. There you go. So in the early days of confocal microscopy, I think we convinced Bio-Rad to loan us a machine as they were in development. So that the, as is common, they could, in those days, they were the only ones fielding the MRC output, which came from Sydney Brenner's lab, essentially when they were trying to do mapping of sea elegance they did serial section EM as you know, but they also were trying to figure out how to make the light microscope work better. So we had one of those tools very early, and we used our capability from just standard epi fluorescence work we'd done before that, to do multiple labeling multiple antibody labeling. So the one on the left is antibodies and non antibody staining. Again, of the cerebellum, the Purkinje cells are oriented the wrong way, that proper convention, it should be flipped upside down, but that's not it looks better this way. It looks kind of like I would just say radishes rooting or something. The one on the right is looking from the vitreous. So looking through the lens of the eye, not really, it's just looking at the surface, the inner surface of the retina and the blue canals are the vessels. The green stars are astrocytes the pebbles, which look like peas on your plate. Those are the cell bodies of the retinal ganglion cells of which they're about 20,000 and the red tracks. Those are the axons of the ganglion cells streaking across the retina, all trying to get to the exit point and the optic nerve. Those are staying with neurofilament. I just like these because they're colorful. And since how we just say, I've always stuck up for the non neuronal cells, the Glea, at least this was one image that I like to show has glea. They are

Peter O'Toole (00:36:00):
Visually, Sorry. They are visually bevy striking. And just finally, cause you sent me this other image as well, which is,

Mark Ellisman (00:36:09):
Yeah, this one is more recent un published. Well, maybe it is published in a paper that one of the fellows in my lab, one of the scientists Matthew Madany, he did again, extending the deep learning based segmentation in this case, it's, again, the cerebellum the red things that looked like you know, red grass. Those are the parallel fibers. And then you see the mitochondria in the dendrites of the [inaudible] cells streaking across the top of the grass there. So again, this is using deep learning based segmentation at a higher I would just say a lower resolution, larger field of view to map sub cellular features and the context of connectomics. So again, just as I explained, this is the kind of merging of chemistry of labeling pushing instruments along, and then having ways of turning complex singings into something that you can actually Block in some useful way, once you look at it.

Peter O'Toole (00:37:27):
And it is that I, I think that as you said, the deep learning to, to, to do that auto segmentation and to do it and this amazing how fast that's come on. If we go back, I don't know. I can't remember the first time I met you. It was taking weeks months to segment an image,

Mark Ellisman (00:37:45):
Right? Usually somebody who gets very strong arms from tracing,

Peter O'Toole (00:37:51):
And now it is getting better. The auto segmentation, the deep learning is getting better. It's still not. How good would you say the that is now compared to still go manual?

Mark Ellisman (00:38:05):
Well, manually you can never do enough to to, to really, you know, bear on an important question, I think, except something very kind of a toy problem. Yep. I think you have to, unless you have a team of tracers as we did 10, 15 years ago where we, you know, get students, then you have to, you know, make sure before you launch on something that is unknown, that they all produce an adequate level of reliability on something is known. The deep learning has started easier to to test with regard to accuracy on test datasets. The problem is the expense of computing because there's the learning phase, which can be computationally expensive. Yeah. And then there's the actual turning loose, the ultimately trained algorithm on a big enough data set. So how practical is it in terms of, I mean, your key question, there was you know, what's the criteria for having a reliable enough surface. I'm translating it to make numbers that are meaningful. Peter, you know, it's going to depend on the question, right. And if you're, if you're trying, I mean, I, you know, let me segue this to another kind of, you know, where, where we are right now is my microscopists. There's been a real boom in the how it just, I call it Nate nativist, microscopy cryo EM, where the rally cry is. And I don't mean to get in trouble that if it isn't in cryo, it isn't true, right. That the only way that you're seeing something which is believable is if it's based on the inherent difference in the way electrons interact with the atoms that nature is pushed together versus water. And if you're looking at something with stain or enhanced contrast with metals as has been the stock and trade in most STEM or TEM or block face imaging, SCM you're somehow removed from reality by some artifact of fixation or embedding or staining. I prefer to think that the structural biology world is enhanced significantly by cryo, and that we'll learn more about, you know, the foals and the atoms and this sort of thing, but that leaves a big gap in what I call the Mesoscale on up. And so if you address your question from the standpoint of what level of detail is going to make a difference for the question, if it's wiring of the nervous system, local wiring plasticity that occurs minute to minute, hour to hour circadian or estrous cycle, or some slow slower cycle aging would be a good one. You need to use Cryo EM to validate the dimensions. You need something that's going to give you enough samples with high enough throughput at, you know, not molecular or atomic resolution, but at cellular resolution. And you sense, they need to determine the difference between pristine cryo structure in a region of interest and what you get as a consequence of all of the preparative methods. And the truth is if you read the literature, there's not much difference if you fix well, high pressure freeze, free substitute, the dimensions are largely the same as cryo.

Peter O'Toole (00:42:20):
Yeah, I don't, I don't think you've done a disservice to the Cryo EM community. I think they are looking at different challenges as well. And, and just like, arguably, therefore an electron microscopist would say that light microscopy is utterly useless, especially when you look at the fixation processes and everything else. And obviously they all have a place in the world of, of getting higher and higher resolution. And it's the resolution you need to see to answer the question that we probably aiming at. And I think that's what you're alluding to essentially, you don't need that Cryo EM ultimate resolution compared to what you can do with a fixed.

Mark Ellisman (00:43:02):
And it's very hard to do that kind of a high resolution work on a sample that comes from an animal's organ. It's great for isolated molecules or something in culture, potentially, maybe some slice reps that are hypoxia tolerant. But for right now, we know we still know how to take it tissues from animals and have them pretty much in the same state they were when the animal was thinking, breathing, whatever. And do metallization the, you may know [Inaudible] from its very inception in the late eighties, wanted to move away from film. And so I spent a lot of time building high quality cameras with CCD devices that were 90% quantum efficient and lenses because I realized that film was great, but it's going to be too slow. And if you wanted to make a microscope that had, that was a, how would you say a smart peripheral device, which we did back at that time that you can either run remotely or having a feedback loop to self-correct you needed a high quality camera and you needed it to be fast, and ultimately you needed it to replace film in terms of the pixel count and the dynamic range. So we sought to do that and we did it, they were expensive. And eventually we shifted to the risky notion that we could get rid of the scintillator and have direct bombardment detectors work. So we went out on a limb 20 years ago and started and got raised money. We even had the patent, we, we invented the direct detector. I mean, we use technology from efforts at the large Hadron Collider. And then we, we manufactured radiation, hardened CMOs devices that wasn't, we didn't do that because we intended to fuel the Cryo EM community, which we did, right. That was a real enabler for which we're credited, even in the Nobel prize for that activity. It was because I wanted high throughput, high fidelity for you know, work in general. Now we're up to an eight K by eight K detector. We have analytics on the same detector. So I still think there's room to grow. And there's important. There's, it's important to continue to remain competitive and serve more than just the cryo community. We use those detectors in STEM, which is really interesting. You coming from more of the physics can imagine you can use a pixilated detector, just like you'd use multi annular detectors and dark field and simultaneously pick up different elements, right? At least you, you're not going to analyze them like we do in energy loss. But if you only have a few elements that give you different radeye in stem, you know, with the right angular capabilities in the platform. So I think there's a lot of room to grow electron microscopes and, you know, I'm blathering a little bit, but one of my frustrations has always been, you know, having learned enough about electron optics to be dangerous, that we're still working with electron optics that are pretty close to [inaudible] design. And if you really wanted to build a microscope that was ideal for a transmission electron microscope, you'd actually have the optical plan be a bit different. And it's a big investment, isn't it? Yeah. I mean, you'd use a Lorenze microscope to be, you know, where the, the, the specimen was not in the immersion field, that the objective for one. But anyway, I think that there's a lot of opportunity. It's just that it's cash. Even though there's a lot of investment in science Microsoft companies are driven by market. So to actually go back to some fundamentals and come up with an alternative microscope column design usually exceeds the interest of the major manufacturers.

Peter O'Toole (00:47:43):
Yeah. It's a long-term development punt, I guess, but there is, I think even without going back to that, there's a lot coming. There's a lot in the background going on in the R&D side at the moment that it's going to be a lot of excitement coming forward, like to ask some quick fire questions, Mark, are you a PC or a Mac person?

Mark Ellisman (00:48:05):
At least for home and my own work, I'm a Mac person. The laboratory is quite mixed and we were you know, Bax, DMS shifted the Unix ahead of most people back in the day. So we've been pretty agnostic, but I like open source tools.

Peter O'Toole (00:48:29):
Are you an early bird or a night owl

Mark Ellisman (00:48:32):
At this point in my life, I am an early bird. And I get up in the morning, the more sleep you get the better. And that's because of the work I do in Alzheimer's disease, .

Peter O'Toole (00:48:47):
That's good advice. I probably might need a bit more tea or coffee.

Mark Ellisman (00:48:55):
Coffee.

Peter O'Toole (00:48:55):
Beer or wine

Mark Ellisman (00:48:58):
Very little wine, a single malt, but less than I used to.

Peter O'Toole (00:49:04):
Okay. Chocolates,

Mark Ellisman (00:49:06):
Conor, single malt

Peter O'Toole (00:49:09):
Sing malt Okay. Okay. What's your favorite brand? Plug it?

Mark Ellisman (00:49:13):
Smoke and dirty. Okay. So I mean, something a little more exotic than log a full-on, but that's, and this is from having been on a board, reviewing the computer science and Edinburgh twice a year for the better part of a decade

Peter O'Toole (00:49:31):
Lagavulin. You must have had lagavulin that is so, ah, anyway chocolate or cheese,

Mark Ellisman (00:49:39):
Cheese.

Peter O'Toole (00:49:41):
Okay. And what's your favorite food? If you go to a conference and someone was taking you out, what would be the best food that they could serve in front of you?

Mark Ellisman (00:49:51):
Well, in the right season in Belgium or Germany, it would be white asparagus,

Peter O'Toole (00:49:57):
Asparagus. And actually I I've eaten in Belgium with you and they, it was white asparagus season when we were there

Mark Ellisman (00:50:03):
Definitely, I think we might've been in that same convent or whatever. And

Peter O'Toole (00:50:09):
That was with Chris Derren, wasn't it? I think I organized it. And it was actually, was it, it might've been ELMI in Leuven.

Mark Ellisman (00:50:18):
Yeah,

Peter O'Toole (00:50:19):
It was that I remember the asparagus being shown before they cooked it on the, on the wood. They were cooking it on. That was good. What is your, why, what is your worst? What do you least like food wise? Is there anything you really dislike food wise?

Mark Ellisman (00:50:38):
No. I'm pretty omnivorous.

Peter O'Toole (00:50:41):
Okay. See, see, see, I really don't like seafood and shrimps and I'm that same meal. One of the courses were shrimps because I got really, he'd gone from this to this, but anyway, hopefully Chris, won't listen to watch this to realize I really found it quite difficult to eat. What about you, what'd you do I do to watch TV? Would you either read a book or watch TV?

Mark Ellisman (00:51:07):
Most of my reading is outside of my main interest. I have scientific hobbies or I spend a lot of time since I've been 19 months here in our little castle. How would you say I've been doing most of the maintenance here? So, okay. I read how I just say and listened to a lot of YouTube on how to deal with stuff. But in terms of what I watch, I would say that Mary and I spend more time looking at documentaries or Nova or no, I've, I've, I've quite a hobby in Marine microbial ecology these days. So I read in that area. Okay. Origin of the eukaryotic cell is a fascination.

Peter O'Toole (00:52:03):
I, I presume there's no trashy TV, a secret vice in. I remember talking to Richard Henderson and he mentioned breaking bad is what he got into actually ever locked down. His big thing.

Mark Ellisman (00:52:16):
I can see Richard breaking bad at some point.

Peter O'Toole (00:52:21):
He's a chemist. I'll give you, I guess.

Mark Ellisman (00:52:25):
I mean, if he went to cell biology, that would be breaking bad for him.

Peter O'Toole (00:52:29):
That's okay into Bible a Bible. That's not go there. I would ask you what your favorite movie is. I'm going to ask you. I'm not going to second, guess it what's your favorite movie?

Mark Ellisman (00:52:43):
Prob probably D depending on it would be something like Citizen Kane for the second tography everybody loves that movie. If you're old enough to have probably watched it. And I mean, there are plenty of movies that I would watch as an uplifting, since I've been chasing gophers around the yard, it could be caddy shack for yeah.

Peter O'Toole (00:53:11):
John Candy wasn't I

Mark Ellisman (00:53:15):
Think he was, but it was I, and anyway, the star was the gopher. So,

Peter O'Toole (00:53:24):
So I I should about what's your trashy movie,

Mark Ellisman (00:53:28):
But there you go. You just give me your trashy movie, rather than trashy tv. I'm more into to comedy than say a science fiction.

Peter O'Toole (00:53:39):
I'm amazed. You didn't go for fight club, obviously. You contributed to the,

Mark Ellisman (00:53:46):
We helped them with the introduction to fight club, but I, it was kinda dark and it came out at a dark time, you know, it was delayed and it's launched because of the Columbine you know, mass shooting. And I guess I'm not keen on movies that glorify any form of violence I'm concerned enough about. I would just say mankind's tendency to, in trying anything that would be pugilistic. Okay.

Peter O'Toole (00:54:20):
If I would say so, obviously for those who don't know the opening sequence was heavy influenced by by science, by true science. And you were the, the lead consultant. I don't know what the right word would be, but this was by a game by complete luck. I don't think they came looking for you. You were sitting on a plane, is that correct?

Mark Ellisman (00:54:43):
Yeah, you've heard me, you've heard me introduce it in a lecture. It was I think it was a very specific I was going to Boston from, I think I, I got the, the nonstop from LA. I flipped to LA and hopped on a plane to get to Boston, to give a lecture at Harvard. And I think I've certain I had enough miles to upgrade to business class or first class or whatever. So I'm sitting on the flight the entire way, and I'm being somewhat pressure motivated. I hadn't really put my PowerPoint together until, you know, sitting on the plane. Right? Yep. I'm working on my laptop. They're on the plane. Hoping not to run out of battery the whole time until we land and as we're landing, I mean, we've usually, you can remember the way I tell the story. Uh I was always afraid of getting into conversations with people because they would distract me from what I had set aside as what I was going to accomplish in that three or four hours. Plus you never know when you're going to get into a conversation with somebody who's overriding thing. I would always put my noise, canceling headphones on and stay quite isolated as best I could. But when you hear the landing gear go down, you can usually, you know, you know, it's a time limited conversation. So I put everything away. And the guy next to me says, I see you're into computer graphics. So am I what do you do? And I said, well, I, I do brain research. And he said, well, I saw, I'm sorry. I peeked at your computer. I saw a lot of images of, you know neurons and things like that. Uh and then he revealed that he was from an animation company, a digital domain, and they had a, a job to do for an upcoming movie. And that he had just gotten the academy award for Titanic. A guy's name was Mike Kanfer. And so I was going, oh, you know, four letter word. I probably should have talked to this guy earlier in the flight, learned something from him. So we had this animated conversation. We realized that I could help him. And that so we were both on the same flight going home at the end of the day, the next day. So we agreed. I showed him my seat number. He rearranged his seat. We sat next to each other the whole way home. And we agreed that I would when I, once I got home, send him some images and that I would talk to him, you know, every week or so, as they were trying to put together an animated scene.

Mark Ellisman (00:57:36):
Now, the part that I thought was most curious is that at that time I was working very intensively on the key non neuronal cell, the most abundant cell in the nervous system called the astrocytes. And we had discovered that astrocytes deployed in the brain very differently than had been described for a hundred years for a hundred years. They'd been described as interdigitating based on seeing the silver staining or the antibody staining for the glial, fibrillary acidic protein. The green cells behind me are astrocytes with just that fragmentary staining. But if you were to see them in their entirety, you would realize that each one has a unique territory. They don't interdigitate at all, and we'd made that discovery. So what I was keen to do with this group in Hollywood was to use their resources, to make a movie that I could then use to point out this transformative way of seeing the piece parts in the brain. So I kept feeding him information, not only about the neurons, which they wanted, but the astrocytes, which filled in the space, but in a cobblestoney way. So I was really keen to get his 92nd video to use it, to convince Washington that we should, you know, be able to pursue this funding agencies. So I was very frustrated when they told me that they weren't going to release the movie. They were holding it because of the potential negative feedback they would get, because it was to be released right about the time the Columbine shooting happened. And I said, well, I've got to go to Washington. I have to go lecture at the NIH. Can I at least get a copy of the intro sequence so that I can try and use it to pitch, oh, Professor Ellisman you're going to be so frustrated because we took all the glioma out. We needed a place to put the camera, cause this starts in the synapse, right with exocytosis and pulls back. And in Hollywood you can show the electrical activity by flashing lights. And so for them, it's a pullback from the fast and fine scale chemistry of electrical and chemical behavior to the, the full facial of Edward Norton, as he's about to pull the trigger and remove Brad Pitt from his brain. So I set out from that point to say, how can I do this on the cheap? They'd spent $6 million for that 90 seconds. And so over a number of years, we figured out how to make animations that were more respectful of the balance of cell types in the brain. So the real, we still think it's under appreciated how the non neuronal cells govern the activity of the neuronal cells.

Peter O'Toole (01:00:52):
It is a great story that I,

Mark Ellisman (01:00:55):
It was a fun thing. And you re you realize how powerful that kind of synthetic graphic is.

Peter O'Toole (01:01:05):
We are actually on the hour at the moment, but I have to ask you first, have you got any top tips for anyone starting out today, neurobiology

Mark Ellisman (01:01:19):
The, for the, those who want to avoid feeling like they're in competition all the time, try and ask questions that are legitimate, but two steps beyond everybody around you. So if you feel that you have, have a good idea, because you can do somebody else's experiment better, probably not the best way to approach science. Okay. I asked where people are not and what questions remain deeper and try and pitch at those, even though they might be harder to fund, you're probably going to be happier if you're not feeling like someone's looking over your shoulder, competing with you all the time, if you can carve out questions that are maybe a little bit more adventuresome and ambitious, maybe harder to fund, as I said, but try and avoid the herd. Yeah.

Peter O'Toole (01:02:23):
More excited. I would say actually, maybe not always more difficult can be more exciting. Grant panels like novelty and something more edgy. Sometimes I have to ask, I asked earlier about what you wanted to be when you were a boy you're now no longer a boy. If you could do any job in the world, what would you be?

Mark Ellisman (01:02:50):
Peter? I have to, I have to think about what other people that I think are more brilliant than me said as they were heading to, you know infinity or whatever it is. Someone like Einstein, who, even though I think he took great solace in fiddling with numbers and problem solving in his head. I think he thought that he probably would have been equally satisfied. Had he been a cobbler? And my interpretation of that was that there some satisfaction that comes from the sense of accomplishment, even if it's, if there's some benefit to other people making shoes and the sense of accomplishment, every time you finish a small job. So how would you say, I, I hope to evolve, to feeling satisfied with things that are less grandiose. So that would be painting in the garage without cutting off my fingers kind of thing. Yeah.

Peter O'Toole (01:04:07):
And you come back to that at some point, mark, thank you very much for joining me today. Everyone who's been watching or listening. Thank you. I strongly recommend very closely associated with this. You've got Jeff Lichtman, you've got Lucy Collinson, Richard Henderson, who we've talked about earlier even go look at etch a cell and the work around that as well. This is some great similar work around this area as well. But mark, you've been very entertaining to talk. I,

Mark Ellisman (01:04:41):
I hope clean enough, you know,

Peter O'Toole (01:04:44):
Definitely.

Mark Ellisman (01:04:45):
I, I live in fear of saying something that I don't understand that would cancel me, but

Peter O'Toole (01:04:51):
No, I, I appreciate you still have got all your fingers despite talking about being in the garage and losing fingers.

Mark Ellisman (01:04:58):
Yeah. I do have a few cuts and scratches.

Peter O'Toole (01:05:04):
We need to get some better YouTube channels to watch how to do that. DIY hope. Thank you very much.

Mark Ellisman (01:05:11):
All right. Thank you, Peter.

Intro/Outro (01:05:14):
Thank you for listening to The Microscopists, a Bitesizebio podcast sponsored by Zeiss microscopy to view all audio and video recordings from this series, please visit bitesizebio.com/themicroscopists.