Grigory Tikhomirov | DNA-Based Molecular Manufacturing for Biotech and Electronics

[00:00] I got a faculty position at UC Berkeley two years ago and now I raised about six million

[00:06] in the first two years of my faculty job.

[00:12] So my dream, I have many dreams, but my main dream was to build a new life form.

[00:19] When I encountered death and sickness in my family, I thought could we build a biology

[00:27] where we don't have to die, where we can upgrade parts of our body on the fly and live happier

[00:33] and healthier.

[00:36] And of course, it's very hard to find this idea, especially if you're an undergrad or

[00:41] a graduate student.

[00:42] So I did my undergrad in chemistry, my PhD is in chemistry too.

[00:48] I grew up in Russia and of course, funding is not that great, political situation not

[00:53] So I'm here just across the bridge from San Francisco.

[00:59] And I have a lab now about 20 people.

[01:02] And so the goal of our lab is to fit this entire Intel factory into a test tube.

[01:12] It's not even molecular printer.

[01:14] So actually, maybe printer is not required.

[01:18] You can just build an information into components.

[01:21] And if you look at how nanoscale devices are built these days, through top down approach,

[01:29] which is probably the most sophisticated technology that humanity ever developed.

[01:35] And that's part of many, many devices that we use every day, like computer chips, microfluidics,

[01:44] optical interconnects.

[01:47] But it's somewhat inflexible approach.

[01:50] And if you know the cost of the most recent asml tool that allows you to do extreme UV

[01:55] lithography is about $500 million.

[01:59] And that's many companies like Intel, Samsung already bought five of them, even though only

[02:05] one was built so far.

[02:08] And if you look at this Intel factory and the devices that manufactured there, you can

[02:16] see if you zoom in that they're not super precise.

[02:21] But if you zoom in on the Pond scum near the parking lot of this Intel factory, you can

[02:29] find a totally different approach to nanofabrication, namely bottom up self assembly.

[02:36] A single bacterium about one micron in size can make billions of different biomolecules

[02:44] in a scalable fashion and with much higher precision.

[02:48] For example, every ATPase in our body is made of exactly the same number of atoms,

[02:54] atomically precise manufacturing.

[02:57] And many of you mentioned it today.

[02:59] Moreover, this molecular devices, structures can exhibit behavior, complex behavior when

[03:09] the potential across the membrane increases.

[03:11] This molecular rotor rotates at different speed.

[03:18] And so my dream in high school was to maybe take this biological components and try to

[03:25] build a new life form where we build things in a more rational way.

[03:32] Because if you now try to disrupt this biological system, like for example, I want to change

[03:42] the electric property refractive index of alpha crystalline, which is a component of OI.

[03:49] Unfortunately, I can cause a cancer too, because it's this alpha crystalline is a

[03:55] component of many metabolic pathways.

[03:58] And so you cannot just take it out slightly change it.

[04:03] It changes the whole pathway.

[04:05] So I turned to engineering and I'm an engineering professor at Berkeley.

[04:11] And there, what I like is a modular approach where you can take transistor, resistor and

[04:19] capacitor, put them together in a very complex integrated circuits and build machines, devices

[04:26] that challenge complexity of a human brain in many ways outperform it.

[04:32] Yeah, well, maybe not for much longer.

[04:35] We we can outperform human brain completely.

[04:42] But if you look at the natural designs, there are many really bad designs.

[04:46] We eat and breathe through the same passageway, major choking hazard.

[04:50] You probably can come up with some other examples.

[04:55] Anybody, bad designs in your body in nature?

[04:58] Yeah, blind spot in the eye, Christine says.

[05:03] Yes.

[05:04] What?

[05:10] Good way to put it.

[05:12] So the sloyringial nerve that I used to speak to you right now, that's two feet of wasted

[05:17] wiring, right?

[05:18] Just because we evolved from fish.

[05:20] You know, as an engineer, you would just nip it, reconnect at the top.

[05:23] But since we evolved from fish, it's not possible.

[05:26] Nano scale, rubisco is the laziest possible enzyme.

[05:31] Typically, enzymes convert thousands of molecules per minute, per second.

[05:36] Rubisco makes maybe two molecules of oxygen from CO2.

[05:41] Sometimes it grabs oxygen and burns glucose.

[05:45] So I said, could we maybe combine the strengths of rational top-down engineering of,

[05:52] let's say, electrical engineering, nanofabrication, and bottom-up self-assembly,

[05:59] something like this, to build a new nanotechnology.

[06:05] And I think it's possible.

[06:07] And I'm not sure what's the best technology, but what I'm using right now is DNA nanotechnology.

[06:12] And many of you know what DNA is, stack of bases, very simple rules, ATGC.

[06:18] You can consider it one and zero.

[06:20] And you can use those rules to build very complex molecules like DNA origami

[06:27] that Adam and some other people already referred to.

[06:31] It's really an amazing process that allows you to assemble, for example, this.

[06:37] So you use ATGC to bring this long black scaffold together with a bunch of synthetic DNA strands.

[06:47] And they'll magically assemble into a structure like this smiley face

[06:52] or map of North America.

[06:53] So basically, it's kind of 2D positioning, no printer required.

[07:00] So what's revolutionary here is you suddenly doubt molecules with the programming language.

[07:06] By programming sequence of these molecules, you can make them fold into any shapes.

[07:13] And I have to say, go Beres, because I'm from Berkeley.

[07:18] And so single DNA origami, it's about 100 by 100 nanometers,

[07:23] and it's big enough to make a single transistor.

[07:26] But if you want to make something more complex like this photoreceptor circuit,

[07:30] we need a much larger breadboard, one minute left.

[07:33] So I'm going to tell you a bit my work as a postdoc.

[07:38] I tried to make a single DNA origami.

[07:41] Postdoc, I tried to make nanoscale monoliths.

[07:47] And it's not Da Vinci, but we can make trillions of this in a test tube.

[07:52] It's atomic force microscope.

[07:55] And I'm just going to quickly switch to tell you what my lab is doing right now.

[08:01] We're trying to bring in materials with interest in physics and use DNA as information bearing molecule.

[08:08] So you can see it's kind of 3D printing.

[08:11] But in a massively parallel way.

[08:13] And there is always the question, what's the first killer app?

[08:16] What would convince this country or bigger scientific community to invest more money?

[08:24] We're not competing with Russia to go to the moon.

[08:28] We're trying to spend taxpayers' money in a sensible way to cure cancer

[08:33] or build some devices that can make profit.

[08:36] So one of these devices is a bear filter.

[08:40] You all have it in your cameras.

[08:41] That kind of allows us to analyze RGB components because you only have three cones in your eye.

[08:49] And so we can assemble this structure in 3D by encoding DNA.

[08:57] By endowing two dielectrics, you only need two different materials.

[09:00] Let's say silicon oxide and titanium dioxide.

[09:03] And just need to attach DNA strands in precise position

[09:07] to make them self-assemble into these complex structures.

[09:11] And we do inverse electromagnetic designs, design software for DNA structure design.

[09:18] And we self-assemble them.

[09:20] And this is an example of how this kind of 3D printer, but a scalable 3D printer,

[09:28] that would allow you to make billions of devices

[09:30] so you can satisfy demands of humanity for this particular component quickly.

[09:37] And as a very first PhD student who joined my lab, he played hockey for MIT.

[09:45] And he got concussion.

[09:47] And since then, he got interested in neurodegenerative diseases.

[09:53] And so he wanted to find a way to read out single neuron activity with millisecond resolution.

[10:02] But we cannot use optogenetics.

[10:05] Many of you know optogenetics.

[10:07] So typically right now it's done by drilling a hole.

[10:11] So that's electrochemical readout, electrical readout.

[10:14] You drill a hole and you insert this neuropixel electrode

[10:20] that can at most read out 200 neurons at the same time.

[10:25] So he is now developing a way to read out many more neurons in a non-invasive way

[10:31] that uses magnetic field that, of course, can go very well through the skull.

[10:37] And this, let me conclude because Alison is looking at me.

[10:49] Yeah, this is my lab.

[10:52] Thank you.

[10:57] Yeah, Durham.

[10:59] Okay, questions, comments?

[11:01] Maybe if number one.

[11:03] I will leave it to the, yeah.

[11:07] Well.

[11:09] So interesting talk.

[11:10] The DNA kind of matching.

[11:12] So you're using the complementarity of DNA as the main kind of organizational tool.

[11:18] But of course, in biology, so DNA has multiple levels of organization,

[11:22] epigenetics essentially.

[11:23] So you can have small scale around a histone or you can have macro domains and such.

[11:30] So has any of that been explored in kind of the evolved kind of macro assembly properties

[11:36] or is it still right now just the small scale interaction that's still being used?

[11:40] Yeah, great question.

[11:41] We also use multiple scales.

[11:44] We do things in a hierarchical way.

[11:47] So ideally what you want to do is to mix thousands of components and they recognize each other.

[11:54] But turns out the yields are pretty low.

[11:57] So what we eventually ended up using is you make a system that, for example,

[12:03] assembles from four building blocks into one building block.

[12:07] You use four of those bigger building blocks to make even bigger and so on.

[12:12] So called fractal or hierarchical assembly.

[12:15] And our body is actually hierarchical.

[12:16] We go from molecules to cells to tissues to organs.

[12:22] What is guiding the larger scale interactions though?

[12:25] I mean fractal, yeah, you just repeat the pattern over and over again.

[12:28] But what is there a greater kind of than interaction?

[12:31] Yeah, we try to keep the same rules.

[12:34] We use DNA sticky ends only for each step.

[12:37] Okay, it's still the sticky end.

[12:38] It's the sequence of ETGCs.

[12:42] Thank you so much.