1,102 words · auto-generated from the episode video
49:41>> So in a new paper out in science from both Harvard and the Broad Institute of MIT is this new genetically encoded device for transcriptto storage information storage >> in mamalian cells. This this story is super interesting. >> Yeah, this is this is pretty incredible because I think it is going to dare I say if this is what it purports it to be, this might be something that is Nobel Prizeworthy. >> Okay. Wow. That's >> I'm going to start there. There's something called the measurement problem in biology. In physics, we have the measurement problem, right? Which is like you measure something and then that destroys the system. Well, in biology,
50:23we've got something very similar, which is the destructive nature of observation. If I want to measure what is happening inside of a cell, I basically have to kill the cell, get all of its contents out, and then do RNA sequencing to figure out what kind of RNA is inside and all that other kind of stuff. So, we can never observe the same cell twice. And what we have to do is we have to infer biological dynamics using static snapshots of stuff. >> You got to break down and then you got to break down here and then you got to sort of connect the dots in time. That's I actually did not know that. That's interesting. >> If you think about it, that kind of makes sense, right? How am I going to observe a cell? Well, I got to break all of its contents. I got to get all the
51:03proteins and then sequence the proteins. Oh, there's this much amount of this protein. There's this much amount of this. This is the mRNA profile. This that which is the transcriptto. So, it's like super nonlinear. And if you have stuff like cancer and embryionic development which is nonlinear and non-erotic these systems you can't really piece together a time sequence that confidently. >> You can't really understand what's happening in time. Uh you can understand what's happening in snapshots of time. >> Exactly. So there's a lot of information loss. Right. There's cellular decisions that happen let's say now. Yes. And then >> the result of that cellular decision is only going to come forward in a few days. >> Yes. Right? So, how do I connect those
51:46two things? And that the existing solutions are something like DNA editing. In 2014, there was this crisper based tape recorder, >> okay, >> that [clears throat] what you could do is you could say, okay, I want to watch these specific genes and how they're transcribed. And there was a tape recorder that you could literally like be like, oh, okay, this thing was transcribed, but this time I'm going to like keep it and save it in some sort of memory. But you had to choose beforehand what to record. Uh it wasn't okay. >> Right. Yeah. And and there's only a few events that you could record before the memory went out. >> Okay. And that's where the breakthrough comes in here. It's called time vaults. Okay. Genetically encoded molecular archives. It's a time capsule for st
52:28cells that stores the secret experiences of their past. >> Very very cool. And it's repurposing something called the vault particle to physically encapsulate mRNA and store it as the cell just goes about its daily business. >> Well, you created almost a cellular backup system like time machine on your Mac. >> Yeah, ex dude. It's exactly that. It's a time machine on your Mac, but now at the cellular level. >> That's incredible. >> It's insane. >> No, that's incredible. Especially given before we were only getting snapped like the the order the the scale of of difference of capability. >> Yeah. Okay. It's >> it's very cool. Okay. So, let's talk about like why this is important, right?
53:08Yeah. So, in physics, we like to think of systems as state vectors. Like all of the particles in a room are going to have position and momenta. And if we know those position and momenta, then we'll know everything about that system. Well, in biology, the state vector is really something called the transcryto, which is the mRNA counts. There's about 20,000 genes that are getting expressed in the human body. And if you could count which genes and how much of that gene is getting expressed in a cell, you could have something like a state vector for that cell. It could be like what is the cell state in >> the the cells in our skin will have a bunch of genes that express for the
53:49proteins that are responsible for our skin. >> The cells in our blood will have the blood genes expressed. The DNA is the same but how much of which gene is expressed is what matters in terms of what that cell is doing. >> It's it's what differentiates how cells be have this multi have these different functions. >> Exactly. Yeah. And so in nature you can think of DNA as like a stable read only memory. So that's your ROM and then your mRNA is a volatile random access memory right that's your RAM. Now mRNA has a halflife that's about 5 to9 hours which which means in about every 5 to9 hours half of the amount of mRNA that was there is going to be gone. It's going to
54:30be degraded. And this is a feature and not a bug because the cell can then quickly respond to an environment and forget what it was doing 5 hours ago. You want to be able to, you know, quickly change what you're doing in case there's some outside stress that comes along. Right. >> It's sort of this adaptive property. >> Exactly. And there were ways that we could tag the new mRNA to to be like, okay, I want I want to check how much of this new mRNA is going through. You could tag it with something called for thiouodine, which is a specific type of nucleotide that goes into the mRNA and it kind of saves stuff. But that signal was limited by about 17 hours. So it's like undetectable after 24 hours. So your memory was only really 24 hours
55:12because you've got these things called exoomes in your cell that are basically the trash collectors and they collect all the trash and then they they put it out. Okay, so okay, how do we how do we