Cheaper microscope could bring protein mapping technique to the masses

While that's true, it'd be exacerbated by a few things (in theory):

1) These instruments would still need to generate income to cover service costs.

2) Income generated per microscope would be reduced because of increased competition lowering beam time prices dramatically, so downtime is still very bad for microscope facilities.

3) More microscopes spread out over a wider geographical area means more service engineers needed (something I've experienced first-hand being in a state with only 3 microscopes, Thermo has been entirely unwilling to place a service engineer here because they can't cover the costs with just 3 microscopes, despite making probably ~1 million/year in service contracts).

In general, I think reduced costs and increased accessibility are a very good thing, but when VPRs go to do the math on these, I think they still don't make a lot of sense.

Never, ever be surprised by the ability of people with large amounts of money to buy expensive toys (err, tools) and then go cheap on all the supporting infrastructure. Like, $1M on a scope, which then runs at about 1/10th capacity because the network is too slow to drain the SSD.

I call this the "You bought a Ferrari to drive on 101 at 10MPH when you really needed a fleet of trucks" problem

Generally most instrument vendors scale the service contract with the cost of the instrument, often around (very roughly) 10% of list price per year. So in a sense a cheaper instrument gets cheaper service. I'm not sure if you'll be sitting down longer (that often depends on how annoying or important you are) or just that parts are going to be cheaper and fewer things that can break. Some stuff is very reliable and you never really get your monies worth from the service contract and some stuff (like electron microscopes) break often.

> I wonder how much potential there is to make them a lot more reliable and easier to use. Other kinds of instruments in the lab have gotten much more reliable and much easier to use over time, though this is not a particularly quick process. But I suspect this depends a lot on how well you can isolate the fragile parts of the equipment from the users.

Going to 100 kV makes everything a little more forgiving in theory. SEMs that operate in the 10 kV range can be had for a full order of magnitude cheaper, although that price probably scales a bit with achievable resolution. But for an example, the 200 kV/300 kV microscopes can't change temperature by +- 1°F over a 24 hour period without the lenses going out of alignment (presumably because of the resistance change in the electromagnets).

>Regarding the last paragraph, it's incredibly frustrating if you see time on expensive instruments wasted because there aren't enough experts around and people have to try and figure out stuff for themselves. But it seems that it's almost always easier to buy another expensive instrument compared to hiring an expert on a permanent position.

The NIH and NSF have been quite willing to provide money to purchase microscopes. They are far less willing to provide money to cover annual operational expenses.

I'm just another grad student doing something between comp bio and structure so take this all with a grain of salt, but I'd read about cryo-ET, FIB milling, structural proteomics, and spatial proteomics (Matthias Mann's group's work comes to mind as an example of some of the wildest stuff people can do). Plenty of groups are working on what you describe but the details are tough. To give a non-technical description of the problem, most structure techniques rely on either explicit or implicit averaging to get good resolution. Small crystals or particle numbers, conformational heterogeneity, and beam-induced radiation damage are big resolution killers. People have figured out all kinds of tricks that amount to purifying a ton of protein, making sure it crystallizes well/is relatively homogenous with a good distribution of ways it sticks to a grid, and minimizing sample beam exposure during diffraction/SPA, which require relatively little, anyway. With cryo-ET, though, you aren't taking a ton of micrographs of a ton of particles or quickly zapping (well, you're rotating it and exposing it for some time under a little jet of super cold air, but hey, it works) a crystal that diffracted precisely because you got a large, relatively structurally homogenous lattice, you're taking a series of micrographs of the same cell, which is structurally probably way more different from other cells than one protein conformation is from another (and that’s if you culture/sort your cells ahead of time instead of working with tissue). Meanwhile, you're frying your sample from the prolonged exposure time a tilt series requires and not even getting a full tilt series because you can't rotate the darn sample holder 360 degrees (though people train CNNs to get around the missing wedge problem now). Once you've reconstructed a structure from your tomographs, how are you going to assign the blobs and squiggles you think you see in your gray static? You either need to know what you're looking ahead of time which requires pre-existing structures and good guesses on location (presumably reasonable to do with something like actin or microtubules, pretty tough with John Doe globular, cytosolic protein that looks like everything else at bad resolution), some kind of labeling (so now you probably need a cryo-CLEM or whatever too or to add something pretty big/distinct which comes with other problems), or you add mass-spec somehow. Oh, and I took the FIB milling for granted when describing this, btw (if you want to look at an entire cell, trying to FIB mill the entire thing is like trying to replace all your walls with glass, so we're back to sampling the right parts/averaging). If you add mass-spec, you’re basically trying to make the smallest laser micro dissection slices ever on a sample that you might have to almost destroy if you want to see more than figments of your imagination. So, hundreds of stupid little problems but people are working on it, and there's work for people with strong computer vision/ML backgrounds.

RNA is a totally different problem since it's often described as pretty "floppy" compared to protein and we don't have that many RNA structures, so presumably it's harder to assign things unless you're looking at well-studied complexes. Aren’t transcriptomics people working on this, though? Maybe MERFISH does kind of what you want just without pipelines for depth assignment yet (or maybe these exist).

if you just want count/location, super resolution techniques (https://www.science.org/doi/10.1126/science.ade2676) and proximity labeling (https://www.biorxiv.org/content/10.1101/2023.10.28.564055v1) may be a good starting point. cryo may be able to help with that if the direct electron detectors get better (as my understanding (and in my experience), cryo-et data is quite noisy). my guess is that multiple techniques will have to be combined and processed with sophisticated computational pipelines to make this a reality. each technique provides some information on the state of the cell, so the computational question becomes if can you figure out how to integrate information between these techniques to get the specific information you want. it may be too early to tackle this, but who knows...

The other answers to your question are quite on point. Of course there are many people interested in this. We aren't there yet, and probably won't ever be (using this technique), but we could begin to paint a picture with lots of correlative methods. Whole cell tomography exists, plus correlated light microscopy, fluorescent tagging of specific molecules, BONCAT, FISH, etc are all techniques that can be combined to give, at least enough of an answer to specific questions that over time we can really build a good picture.

Whole viruses, obviously, have been done and are quite interesting.

AlphaFold is a real game-changer in predicting many protein structures, but its precision in dealing with single residue mutations, particularly in non-standard proteins, isn't a sure bet.

The tool excels because it's been trained on a massive database of known protein structures. It's great at making educated guesses based on that data, but it's not as reliable when it comes to variations that don't have much historical data, like specific mutations at the residue level.

For these finer details, traditional physics-based methods, like molecular dynamics simulations, might offer more insight. They really get into the atomic-level interactions, which can be critical for understanding the subtle effects of amino acid changes.

AlphaFold is likely to identify significant structural changes, but it might not be your go-to for pinpointing smaller, more nuanced shifts.

AlphaFold is basically a pattern matching program on steroids. Don't get me wrong, it's incredible, but if there's not experimental data somewhat describing what that variant would look like, AlphaFold is liable to just make spaghetti out of it.

The Tundra is supposed to be aimed at the same audience as this microscope. It's Thermo's smaller, cheaper, easier to maintain microscope. They are essentially trying to close off the bottom of the market from their competitors (JEOL and Hitachi). They are rightfully concerned that if people start buying 100 kV JEOL and Hitachi microscopes to learn/screen with, they will start preferring those when they go for large data collections, and that will erode the cash king that is the Titan Krios 300 kV TEM.

It's quite new, and there are many many issues with it, from a value standpoint. First of all, its' really not much cheaper than their previous budget 200 kV Glacios, and it's not much cheaper than JEOL's 200 kV offering as well. But there are rather insane compromises that make no sense.

They are saying that you need less space for the microscope, and while that's true, you also need an additional "loading station" that is quite large and necessarily complicated, that you don't need for the 200 kV and 300 kV microscopes. As an example, the loading station is a large desk that needs its own liquid nitrogen source - not needed on other microscopes. It has a computer that loads clipped grids into an autoloading arm - not needed on other microscopes. It uses the same expensive 12-grid cassette, but you can only use I think 5 of the slots in it, and only one at a time - again, not needed on other microscopes, which can load all 12, at the same time.

The microscope needs manual nitrogen filling, you can't just hook up a tank to it. You need to manually open the microscope doors, and pour nitrogen in, which is an awful idea for new users and seriously limits its autonomy. You can't just queue things up an forget about it any longer.

So in terms of ease of use, they've added a ton of complexity, but they haven't really reduced costs all that much. It really seems like a double miss.

Where it gets insidious, though, and Thermo displays their true hand, is on the software side. This is somewhat esoteric to those not in the field, but basically Thermo has always (since they were Phillips, and later FEI), sold their microscopes with a scripting package that you can purchase. You can then use that scripting package to control basically every aspect of the microscope. That lead to the development of open source software such as SerialEM, and Leginon. Those software packages are single-handedly responsible for the automation of data collection that has enabled cryo-EM protein structures. It would not have been possible on Thermo scopes without it, and frankly, Thermo didn't have the foresight to design their software to be capable of the techniques those open source packages employ. i.e. the exploding field of cryo-EM that Thermo capitalizes on by selling many 5 million dollar Krios microscopes every year would never have existed to begin with, if it weren't for the scripting package. And every year, people come up with new, ingenious ways to use that scripting package to do new, innovative things.

Well, now with the Tundra, Thermo has gotten rid of the scripting package. You are stuck using their data collection software. You can't even manually align the microscope anymore - everything is "automated" for "ease of use". Additionally, they've introduced multiple new APIs, some of which you are required to buy separately, if you want access to specific features. Essentially, they are going for two things: artificial market segmentation, and trying to get users locked in on their software from the ground up. It has not been well received, and has severe limitations, and in my opinion, could really use an antitrust investigation.

Are you thinking of scanning electron microscopes? Those will give you surface topology - the electrons don't pass through the sample, but instead backscatter. Platinum and gold are commonly sputtered on top to help with the conductivity.

With Cryo-EM, it's a transmission electron microscope. So the sample is frozen into a very thin (~30ish nm if you're lucky) sheet of vitreous ice, and then shot with electrons. The electrons pass all the way through the ice and the sample, and you get a projection image of whatever was in that ice.