用人工智能治愈所有疾病——马克斯·亚德伯格

但你刚刚告诉我，不是这样的。

But you've just told me, no.

它会持续在整个过程中运动。

It just keeps moving through the whole thing.

是的。

Yeah.

这些确实是极其复杂的动态系统，是的。

These these these are really, really complex dynamical systems Yeah.

你知道，它们由成千上万、数百万甚至数万亿个原子组成，在细胞内以微秒乃至更短的时间尺度展开。

You know, composed of, you know, thousands, millions, trillions of atoms within our cells unfolding over the course of, you know, microseconds and and and beyond.

你所说的这种动态性

And this dynasism that you're talking about

但是什么

But what

这个词是什么？

word is that?

动力学？

Dynasism?

那是个词吗？

That's a word?

那不是一个词吗？

Isn't that a word?

动力学？

Dynasism?

我觉得我

I think I

刚编出来的

just made

这个词。

it up.

错位。

Mism.

哦，动力学。

Oh, dynanism.

不。

No.

动力学。

Dynamism.

动力学。

Dynamism.

谢谢。

Thank you.

谢谢。

Thank you.

我在纠正语法。

I'm correcting grammar.

这是

This is

第一次。

a first.

这是。

This is.

动态性。

Dynamism.

这种动态性。

The dynamism.

同一种恐龙。

The same kind of dinosaur.

动态性。

Dynamism.

不是恐龙。

Not dinosaur.

动龙。

Dynamisaur.

不是恐龙。

Not dinosaur.

恐龙。

Dinosaur.

但你所说的细胞内的动态性。

But the dynamism that you're talking about within the cell.

当你观察我们每一个人时，由于我们每个人都是如此不同，即使有着共同的执行方式和蓝图，我们每个人却都如此不同。

When you look at each one of us, since each one of us is so different, even though there's a general execution and blueprint, we all come out so different.

这是你正在观察和绘制的过程的一部分吗？

Is that part of the process that you are looking at and mapping?

我会对水母说，我们都长得一模一样。

I I would say to a jellyfish, we all look identical.

对。

True.

明白？

Okay?

他们并不是在说，你的肤色略有不同，或者你稍微高一点？

They're not saying, oh, is your skin color slightly different or you're slightly taller?

是的。

Yes.

你是在描述细胞层面的功能。

You're describing functions at a cellular level.

嗯。

Mhmm.

你的工作是理解这一点，还是去探索生物学尚未发现的蛋白质折叠新方式？哦。

Is your job to understand that or is your job to figure out extra ways to fold proteins that maybe biology has yet to even figure out Oh.

这些方式可以解决自然宇宙尚未遇到的问题。

That can then solve problems that we encounter that the natural universe has not.

所以，哦，这是个好问题。

So oh, that was a good question.

我正想说，太多了。

I was about to say so much.

我对

I'm happy with

自己很满意。

yourself.

今天我对自己非常满意。

I'm so happy with myself today.

是的。

Yeah.

这真的很有意思。

This is really interesting.

我们有这些微小的分子机器——蛋白质，我们关注它们的三维结构及其功能，原因有两个。

We we have these little molecular machines, these proteins, and we care about that three d structure and how they work for two reasons.

第一，我们想了解细胞是如何运作的，因为如果这方面出错——这正是疾病发生的情况，对吧。

One, we want to understand, you know, how our cells work because if something goes wrong with that, which is the case for disease Right.

然后我们还想弄清楚，我们应该在哪里介入并开始修复这些问题？

Then we want to understand, okay, where can we or where do we actually need to go in and start fixing that?

对。

Right.

或者我们如何从一开始就阻止它出错。

Or how we can stop it from actually going wrong in the first place.

没错。

Exactly.

没错。

Exactly.

所以这就是其中一点。

So that that's one thing.

然后当我们思考，好吧，我们该如何去修复它？

And then when we think about, okay, how can we go and fix that?

当我们进行药物设计时，实际上是在说：我们能否创造出另一种分子，进入细胞并开始调节这些分子机器？

What we're actually saying when we're doing drug design, we're saying, can we create another molecule that will come into the cell and actually start modulating these molecular machines?

这个药物分子将会附着在那边的蛋白质上，从而导致该蛋白质改变形状，例如，使其不再像平常那样运作，或者阻止它工作，又或者让它更好地工作。

It's going this this drug molecule is going to actually attach to this protein over here, and that's gonna cause this protein to change shape, for example, and so it won't operate how it normally does or so we stop that protein working or we make it work better.

这些就是我们在

These are the sort of things we do in the

让我想起了我们为新冠开发的mRNA疫苗。

reminds me of messenger RNA vaccines that we developed for COVID.

是的。

Yeah.

你知道，我们在药物设计中利用了多种不同的分子机制。

You know, there's so many different types of molecular mechanisms that we take advantage for for drug design.

我是阿利坎·赫马吉，我在Patreon上支持StarTalk。

I'm Alikan Hemraj, and I support StarTalk on Patreon.

这是尼尔·德葛拉司·泰森的StarTalk。

This is StarTalk with Neil deGrasse Tyson.

蛋白质折叠通常是否遵循某种固定模式，你们能够对其进行映射吗？

Are the folding proteins generally following a set pattern in the way that they do fold, and you're able to map them.

当它们错误折叠时，你们就能发现异常，还是我只是重新发明了什么，或者说了些废话？

And when they misfold, that's when you're able to flag that up, or have I just reinvented something or talked rubbish?

那太酷了。

That'd be cool.

不。

No.

不。

No.

你说到点子上了。

You're you're you're you're onto something.

所以，所以

So so

那太好了。

That'd be great.

所以，你知道的，对吧？

So, you know Right?

我的意思是，令人惊叹的是，我们实际上能够预测这些蛋白质是如何折叠的。

Like, the I mean, the the amazing thing is that we can actually, turns out, predict how these proteins fold.

嗯。

Mhmm.

所以它们是，所以

So they are So

你正在模拟这个过程。

you're modeling that process.

我们用深度学习和神经网络来模拟这个过程。

We're modeling that with deep learning, with neural networks.

这就是AlphaFold以及它的所有后续版本所关注的。

That's what AlphaFold, you know, and all its generations are all about.

这意味着，我们只需输入一段氨基酸序列，对这个蛋白质一无所知，就能输出它的三维结构。

And, that means that we can actually just take in a sequence of amino acids, knowing nothing about this protein before, and then get out the three d structure.

而通常，人们要花上几个月甚至几年才能解开这个结构。

And normally, this would take people months, if not years, to work out this

那么，AlphaFold是如何知道一个大分子会如何折叠的呢？

So how is it that AlphaFold knows how a large molecule wants to fold?

再次，它必须以某种方式知道这一点。

Again, the It's gotta know that in some way.

它是从几十万个例子中学习到的。

It it it's it's learned this from a few 100,000 examples.

因此，在过去五十年里，化学家和生物化学家们一直手动解析这些蛋白质结构。

So chemists, biochemists over the last fifty years, they've been working out these protein structures by hands.

他们实际上是

They've been literally

哇。

Wow.

合成蛋白质，将其结晶，然后用X射线照射，观察电子散射，从而解析出蛋白质结构。

Synthesizing protein, crystallizing it, then shooting X rays at this to, like, look at the electron scattering, and from that, you can resolve the protein structure.

这是一个相当困难的过程。

It's it's it's a pretty hard process.

但人们一直在做这件事。

But people have been doing that.

那是你拍摄分子形状的方式，那就是

That's your way to photograph what the shape of the molecule That's

你拍摄现实的方式。

your way to photograph the reality.

通过

With the

用那种方式，基本上就是一台电子显微镜

with that kind of it's basically an electron microscope

在那个层级上。

at that level.

是的。

Yeah.

相似。

Similar.

就像电子散射。

Like electron scattering.

是的。

Yeah.

对。

Yeah.

没错。

Exactly.

因此，过去五十年来，人们一直在做这件事，沉积这些结构。

And so people have been doing that for the last fifty years and depositing these structures.

现在，我们收集了所有这些数据，训练了一个神经网络，仅根据分子的描述输入来预测所有这些数据。

And now we've taken all of that data and trained a neural network to go just from the input of what is this molecule description to try and predict all of that data.

对。

Right.

令人惊叹的是，你可以用过去五十年的数据来训练它。

And the amazing thing is, and this this is really remarkable, is that you can then train this on the last fifty years of data.

这涵盖了大约二十万个蛋白质和生物分子系统，但你似乎可以将其应用到我们所知的蛋白质宇宙、蛋白质组中的所有事物上。

That's a couple of 100,000 protein and biomolecular systems, but you can apply it seemingly to everything we know about in the protein universe, in the proteome.

那么，它有Proteum吗？

Well, it has Proteum?

蛋白质组。

The proteome.

哦，我们喜欢这个。

Oh, we like that.

哦，蛋白质组。

Oh, proteome.

Proteum。

Proteum.

Proteum。

Proteum.

是的。

Yes.

Proteum。

Proteum.

Proteum。

Proteum.

哦，哇。

Oh, wow.

是的。

Yeah.

那么有多准确

So how accurate

AlphaFold和AlphaFold有多准确？

is AlphaFold and AlphaFold?

我们已经进行了第三轮预测，因为AI已经存在一段时间了，正如你所说，你并不是唯一可用的AI工具。

We're on the third iteration with its predictions because AI has been around a little while, as you've already said, and you're not the only AI tool that's out there.

但这个特定工具的准确度如何？

But how accurate is this particular tool?

是的。

Yes.

所以AlphaFold 2

So so AlphaFold two

对。

Right.

是那次巨大的飞跃，让我们开始获得实验级别的准确性

Was that big jump where we started to get experimental level accuracy

嗯。

Mhmm.

仅针对蛋白质，也正是因此获得了诺贝尔奖

For just proteins, and that's what won the Nobel

奖项的推广。

Prize campaign.

与实际实验相比，是的。

Off against empirical experiment Yeah.

基准就是实际的实验室工作本身。

The the benchmark is doing the real lab work itself.

对。

Right.

所以AlphaFold 2达到了那个水平。

So AlphaFold two reached that level.

现在AlphaFold 3将范围从仅蛋白质扩展到包含其他生物分子类型。

Now AlphaFold three expands from just just proteins to incorporate other biomolecular types.

比如蛋白质与其他蛋白质、蛋白质与DNA、RNA以及所谓的小分子，它们都属于邻近区域。

So proteins with other proteins, proteins with DNA, with RNA, with what's called small molecules, which are They go to the neighborhood.

它们开始把这些都混合在一起。

They start mixing all that up.

或者也许不是邻近区域。

Or maybe not the neighborhood.

也许那个邻近区域得到了一些升级。

Maybe that neighborhood gets a little bit of an upgrade.

不。

No.

那就是你创造出超人类的时候。

That's that's when you make superhuman.

我可能会试着

I might just try

成为终结者。

to be the Terminator.

它不会是终结者。

It's not going to be the Terminator.

它会是超人类。

It's going to be the Superhuman.

然后它们会像我们一样，俯视我们，心想：我们为什么还需要你们这些人？

And then they're going to be like us and they're going to look down on us and go, you know, why do we need you guys?

就这样了。

And that's it.

所以，无论如何，你能预测这些结构，你有没有实际根据这些建模预测来合成蛋白质？

So anyway, you're able to predict these and has have you actually taken any of the modeled predictions and made the proteins?

是的。

Yeah.

哦，对。

Oh, yeah.

或者告诉我们，你期望这些成果如何催生全新且创新的药物。

Or or tell us where you expect these to lead to new and innovative drugs.

否则，这仅仅是个拼图游戏。

Cause otherwise, it's just a puzzle exercise.

是的。

Yeah.

这就像一套很棒的乐高积木。

It's a great LEGO set.

我们

We

想要

want the

让客人享受这一点。

guests to enjoy this.

是的。

Yeah.

天啊，真是这样。

It's like, oh my God.

这个乐高套装要多少钱？

How much that Lego set costs?

只要100亿美元。抱歉，你请说。

Only $10,000,000,000 Sorry, go ahead.

是的。

Yeah.

对。

Yeah.

所以，如果你针对某种特定疾病，我们其实可以通过调节某种特定蛋白质来解决这种疾病。

So so so if you take a particular disease and we we, you know, we we we we that we can actually, you know, solve this disease by modulating a particular protein.

嗯嗯。

Mhmm.

问题是，我们该如何做到这一点。

The question is how we do that.

所以我们设计了一种药物分子，希望它能以某种特定方式与这种蛋白质结合。

So we design a drug molecule, and we want it to fit to this protein in a certain way.

哦。

Oh.

因此，传统上，你只能靠猜测。

And so this is where traditionally, you would have to actually either just guess.

对。

Right.

或者去

Or go into

然后对每一种组合进行结晶、拍照，看看是否有效。

And the create crystallize each one of those combinations and then photograph them and see if it worked.

但现在你可以建模，AI能在一分钟内完成上千次这样的模拟。

But now you can model it, and the AI can do a thousand of those in, like, a minute.

这不是就是那个……

Isn't this what There

你懂的。

you go.

靶向蛋白吗？

Target target proteins?

是的。

Yes.

哇。

Wow.

对。

Yeah.

所以如果你知道某个特定的靶向蛋白，是不是就可以用它去筛选一批药物，然后想，嗯，这个药物A对这个蛋白效果更好，或者可能是药物D，或者其他字母代表的药物？

So if you know you've got a certain target protein, do you not then run that against a list of drugs and think, yeah, this one, drug a works better with this or maybe it's drug d or whichever letter of the alphabet you're on?

现在我们变成了侦探，AlphaFold 3 产生了多少线索和答案，还是我们仍在努力探索

And now we become the sort of detective, and has this AlphaFold three produced how many clues and how many answers, or are we still grappling

与是的。

with Yeah.

与其试图找出药物，AI 实际上已经为你找到了药物。

Instead of trying to figure out the drug, the AI actually figures out the drug for you.

药物发现。

Drug discovery.

没错，正是如此。

Well, exactly.

你知道，一旦让那个人说下去，我们俩就都听到了

Know, once you let the guy speak, we both hear the

你们两个。

two of you.

我们自己正在摸索整个这个行业，加里。

We're figuring out this whole industry ourselves, Gary.

是的。

Yes.

对。

Yeah.

你知道，这正是它的发展方向。

You know, this is exactly where it's going.

没错。

So Right.

我们可以开始理性地设计这些药物。

We can start actually rationally designing these drugs.

对。

Right.

传统上，你会取数百万个随机分子，把它们扔到这些蛋白质上，看看哪些能结合上。

Traditionally, you would take, you know, a million random molecules and you would just throw them at these proteins and see what sticks.

而许多药物正是这样历史上被发现的。

And and and that's how so many drugs have been created historically.

是的。

Yeah.

如果你再往回追溯，你就是在泥巴里筛选这些分子。

You you go back further and you're sifting through mud to find these sort of molecules.

这就是为什么针对各种问题的药物成功率一直这么低吗？

Is this why there's been such a a low percentage of success rates with the sort of drugs for whatever the problem is?

这是部分原因，因为我们并不完全理解这些分子是如何起作用的。

That's part of it because, you know, we we we we don't necessarily understand how these molecules are working.

但有了像AlphaFold 3这样的技术，你可以把分子和目标蛋白输入系统、输入神经网络，然后得到它的三维结构。

But with something like AlphaFold three, you can, you know, put the molecule, put the target protein into the system, into the into the neural network, and you get out the three d structure.

作为化学家，你可以开始理解：这个小分子药物是如何调节这个蛋白质的？

And as a chemist, you can start to understand, okay, how is it this small molecule drug modulating this protein?

现在，问题仍然是：你该如何首先找到那个适合这个蛋白质的小分子呢？

Now, still the problem is, well, how do you find that small molecule in the first place that's gonna be good for this protein?

据估计，可能存在大约10的60次方种潜在的药物样分子。

You know, it's estimated there's, like, 10 to the power of 60 possible drug like molecules out there.

那就是1后面跟着60个零。

That's, you know, 10 with 60 zeros.

所以即使人们知道

So even people know

10的60次方，是的。

what 10 to the 60 Yeah.

是的。

Yeah.

对。

Yeah.

好的。

Okay.

别没礼貌。

So Don't be rude.

即使你有完美的AlphaFold

Even if you had the perfect alpha fold

嗯。

Mhmm.

你必须在10的60次方个分子上运行这个计算，这在计算上是不可能的。

You'd have to run that across 10 to the power 60 molecules, which is just computationally impossible.

这是不可行的。

It's unfeasible.

对。

Right.

直到量子计算出现。

Until quantum computing.

直到那时，那么接下来会怎样呢？

Until And so then what

我们需要的是一个我们称之为生成模型或智能体的东西，它能够实际探索这个空间，理解整个分子空间，并为你设计出分子结构。

we need is is something that we call a generative model or an agent which is able to actually search through that space, understand that entire molecular space, and come up with molecular designs for you.

哦，因为10的

Oh, because the 10 to

如果只是随机尝试的话，10的60次方就是如此。

the 60 is if you just did it randomly.

对。

Right.

对。

Right.

但是

But

这纯粹是碰运气。

That's just throwing it at anyway.

对。

Right.

对。

Right.

如果你不随机尝试，那就可以，是的。

If you don't do it randomly, then you can Yeah.

但随机性才是当前的前沿方法。

But but but randomly is the state of the art method.

人们就是这么做的。

That that's how people do it.

这就是人们目前的做法。

It's how people currently do it.

这就是人们目前的做法。哦，你称它为前沿技术。

It's how people currently do Oh, you well, he called it state of the art.

你就是前沿技术。

You're the state of the art.

对。

Right.

谢谢。

Thank you.

让我们在这里正确地使用这个词。

Let's let's use the word properly here.

那么，如果蛋白质实际向左折叠，而你预测它应该向右折叠，该怎么办呢？

So what if what if the protein turns left when you've predict when you've mapped it to turn right?

这是否就是AlphaFold 3也会遇到问题的情况？

Is that when we have issues that even AlphaFold three has a problem with?

没错。

Exactly.

归根结底，这些模型并不完美。

These are, not perfect models at the end of the day.

它们非常非常准确，但仍然会犯一些错误。

They're very, very accurate, but they will make some mistakes.

所以，目前你仍然需要偶尔进入实验室，但你需要做的实验工作量已经少了很多。

So you still do, currently, need to go into the lab occasionally, but the amount of lab work you have to do is so much less.

对。

Right.

而且你常常能发现这些模型表现得特别特别好的分子空间区域。

Then and and you and often you can find the area of molecular space where these these models work really, really well.

然后我们会在稍后的时间去实验室。

And we then go out to the lab later later down the line.

我们会对这些物质进行结晶，并看到，是的，这正是模型预测的完美映射。

We crystallize these things and we see, yeah, like this is a perfect mapping of what the model predicted.

所以回到之前的一个观点。

And so back to an earlier point.

对。

Yeah.

在过去，比如上个月，制药公司，大型药企，会花费数百万美元，也许不到十亿，但高达数亿美元来开发一种药物。

In the old days, like last month, you, the the pharmaceutical companies, big pharma, would spend millions, maybe not quite a billion, hundreds of millions of dollars developing a drug.

我们认为，撇开定价可能存在的滥用问题，确实存在这样的事实：第一粒药片的成本

We think that holding aside what might be abuses of pricing, the fact that there's some truth to this first pill

嗯。

Mhmm.

高达五千万美元。

Cost $50,000,000.

第二粒药的成本是10美分。

The second pill cost 10¢

嗯。

Mhmm.

因为他们必须进行研究才能找到第一粒药的配方。

Because they had to research to get the formula for that first pill.

如果你缩小了搜索范围，那么开发第一粒药的成本就可以大幅降低。

If you have narrowed the search space, then the cost of developing that first pill can be manifold smaller.

平均而言，研发一种新药的成本是30亿美元。

It costs on average $3,000,000,000 to create a new drug.

哇。

Wow.

对吧？

Right?

这是平均值。

That's that's that's that's on average.

是的。

Yeah.

所以，这是一个所以它

And so, there's a So it

当我提到一亿时，你低估了。

was low when I said a 100,000,000 You were lowballing.

低估了，好吧。

Lowballing Okay.

所以，这是一个巨大的机会，可以彻底改变成本和速度。

So this is a massive opportunity to, like, completely change just like the cost, the speed.

所以商业模式。

So the business model.

是的。

Yeah.

而我们的商业模式正如我们所做的那样是

And the business model as we do Like is

是专有的吗？

it proprietary?

所以，我真正的想法是，这里才是你能实现革命的地方。

So here's my real because here's where you would revolutionize.

如果我发明了它，而我是A公司，对吧？

So if I come up with it and I'm company A, right?

它属于我，我可以决定一切。

It's mine and I get to determine everything.

如果你是一家AI公司，只是做这个以便出售，那么它就属于你。

If you're an AI company and you're just doing this, okay, so that you can sell it, then it's yours.

哪一个会让消费者的成本更低？

Which one will make prices lower for the consumer?

我们的目标是真正重新定义药物设计的方式，让成本大幅降低。

Our goal is like really redefine this this way you do drug design so it becomes so much cheaper.

我们拥有了更多潜在药物和化学物质的资源，这确实改变了商业模式，也改变了这个领域的经济格局。

We have so much more abundance of potential drugs and chemical matter that it really does change the business model, and it changes the economics of the space.

所以你实际上可以彻底改变药物制造的成本。

So you you can actually revolutionize the the cost of making drugs.

是的。

Yeah.

对。

Yeah.

这正是我们正在走向的方向。

That that's that's where we're going.

那就是我们的目标。

That's where we're going.

好的。

Alright.

我感到满意。

One of I'm satisfied.

AlphaFold下一步的举措之一，无论通过当前版本还是下一个迭代，都将探究导致蛋白质错误折叠的原因和驱动因素，以便你能提前介入整个过程。

Is one of the next steps with AlphaFold, whichever it's through or maybe the next iteration or so, going to investigate why and what drives the misfolding of a protein so as you can kinda get ahead of even the story

哦。

Oh.

这种事情的发生。

Of that happening.

哇。

Wow.

实际上，蛋白质的错误折叠是另一回事。

So so, actually, the misfolding of a protein is another thing.

这会导致某些类型的疾病。

That that's what causes some types of disease

是的。

Yeah.

当你体内发生基因突变，DNA中的突变会改变蛋白质中的某个特定氨基酸，导致它无法正常折叠。

Where you'll have a genetic mutation, a mutation in your DNA which will change a particular amino acid in that protein, and so it doesn't fold the normal way it should fold.

对。

Right.

因此它无法像正常的分子机器那样发挥作用。

And so it doesn't function as as it normally should as a molecular machine.

因此，像AlphaFold这样的工具可以帮助我们理解哪些突变会导致蛋白质错误折叠，这些突变被称为错义突变。

And so things like AlphaFold can help us understand what are those mutations that cause misfolding, and they're called missense mutations.

对。

Right.

而这些可能成为潜在的药物靶点。

And then, you know, these could be potential drug targets.

因此，我们可以考虑设计一些分子来缓解这种情况。

So we could, you know, think about molecules that could, you know, mitigate against that.

如果我理解得没错，当你看PDR时，它有这么厚，PDR是什么？

If I understand correctly, if you look at the the PDR, it's this thick and What's PDR?

医师案头参考手册。

The physician's desk reference.

谢谢。

Thank you.

它有这么厚。

It's this thick.

所以它和老式曼哈顿电话簿一样厚。

So it's the size of an old style Manhattan phone book.

明白吗？

Okay?

它非常厚，有好几英寸宽。

It's very thick, multiple inches across.

里面充满了医生可以开处方的现有药物。

And it's chock full of existing medicines available to the doctor to prescribe.

真的所有这些药物都是通过化学作用与患者互动，而不是通过蛋白质折叠吗？

Is it true that 100% of those medicines are interacting with the patient chemically rather than through protein folding?

如果是这样，那是否意味着当蛋白质错误折叠时，我们无法用任何折叠算法来应对，只能依靠化学药物来帮助身体应对这种影响？

So that if that's the case, does that mean that where proteins misfold, we can't combat it with any kind of folding algorithm, we just prescribe chemistry for your body to handle the impact of that.

我刚才说的对吗？

Is that did I say any of that right?

我认为这确实是大多数药物的情况。

I think that that that is the majority of drugs.

是的，没错。

They are Yes.

它们是我们服用的化学物质。

They they are chemicals that we take.

我们就在那里。

We are there.

我们，你知道的，服用。

We we, you know, take

药片。

the pills.

无法修复折叠问题。

Not gonna fix the fold.

它只是缓解症状。

It's gonna treat the symptoms

对的。

Right.

针对发生的错误折叠。

Of the of the misfolding that happened.

我们并没有改变蛋白质的突变。

We're we're not we're not changing the mutations of the proteins.

那可能是一种基因疗法。

That that could be something like gene therapy.

嗯。

Mhmm.

对吧？

Right?

但这些是进入体内并附着在这些蛋白质上的化学物质，以某种方式减轻错误折叠，或者改变相互作用界面，改变这些分子机器的工作方式。

But these are these are chemicals that come in and will attach themselves to these proteins and somehow mitigate the, you know, something like a misfold or it'll change an interface, change how these molecular machines work.

这太神奇了。

That's wild.

目前Isomorphic实验室特别关注哪种疾病，还是这是一个更广泛的应用？

There a particular disease isomorphic labs are focusing upon right now, or is this a more of a broad spectrum?

我们是针对蛋白质逐个挑选某些目标，还是真的只关注某一种特定的？

Let's do let's let's go for proteins and cherry pick out certain things, or will we really looking at one particular?

我们正在开发的技术是非常非常通用的。

The technology we're creating is really, really general.

我们希望将这个药物设计引擎

We wanna be able to apply this drug design engine

好的。

Alright.

应用于任何蛋白质、任何靶点、任何我们遇到的疾病领域。

On any protein, any target, any disease area that comes our way.

话虽如此，作为一家公司，实际上你需要聚焦于某个特定领域，我们目前主要关注癌症方面。

Now saying that, you know, practically as a company and and that you wanna focus on a particular area, we're focusing at the moment on, you know, a lot on cancer

当然。

Of course.

还有免疫学。

And a lot on immunology.

当然。

Of course.

我太重要了。

I'm too biggies.

我太重要了。

I'm too biggies.

是的。

Yeah.

是的。

Yeah.

而其中两个最符合你想要实现的目标的，

And the two that probably lend themselves best to what you're trying to accomplish,

实际上。

actually.

好的。

Okay.

现在观看并收听这段内容的每个人都会想让我问的问题是：你进展得怎么样？

The question everyone's gonna want me to ask right now that's watching this and listening to this is, how are getting on?

老实说，情况非常好。

Would I mean, it it it it's going really well, to be honest.

嗯。

Mhmm.

我们正看到这些算法真正改变了我们进行药物设计的方式。

We're seeing these algorithms actually, you know, change the way that we're able to do drug design.

我们能够发现新的分子结构。

We're able to discover Mhmm.

你知道，针对一些人们已经研究了十多年甚至更久的靶点，发现完全新颖的化学物质。

You know, completely novel chemical matter against some of these targets that, you know, people have been working on for, you know, even over a decade.

对吧？

Right?

所以，非常困难的东西，却取得了惊人的进展。

So really, really hard stuff, making amazing progress.

是的。

Yeah.

公司还处于非常早期的阶段，但确实如此。

Still still really early in the company, but Yeah.

这太令人兴奋了。

It's super exciting.

你们已经送过东西去拍照了吗？

And have you sent anything to be photographed yet?

我们确实送了一些东西去做分子成像。

We we we we send some things for for molecular photographs.

是的。

Yeah.

好的。

Okay.

是的。

Yeah.

对。

Yeah.

我知道你不能谈论这个。

I know you're not allowed to talk about it.

我知道。

I know.

我明白。

I get it.

但我们都很开心。

But we're all very happy.

好的。

Okay.

就是这样。

There you go.

那是听。

That's listen.

我明白你的意思。

I'm I'm with you.

我懂你表达的意思。

I'm picking up what you're putting down.

这很棒。

That's cool.

是的。

Yeah.

但如果你把这项工作视为基础研究，那么你发表成果时，就像他们发表DNA分子图像那样，让大家知道它是双螺旋结构。

But so if you if you I see this work as fundamental research so that you publish a result, you publish the image as they published the image of the DNA molecule to know that it was a double helix.

没错。

Exactly.

在那一刻，它就成为了公共知识。

That becomes public knowledge at that point.

所以，一旦你们发布了AlphaFold 3的蓝图，任何拥有工具和访问权限的公司都能使用它。

So someone with tools, access to AlphaFold three, would any company have access to this once you have published the blueprint for it.

在药物设计中，这些蓝图通常会出现在专利中。

In in drug design, often these blueprints come out in the patents.

所以，当你准备进入临床试验时，你需要为这些分子申请专利。

So when when you're when you're gonna go into clinical trial, you need to patent these molecules.

在这些专利中，你会包含大量关于分子及其化学式的信息。

And in those patents, you'll have a lot of data around the molecules, the formula

这正是我之前提到的。

That's I was talking about earlier.

是的。

Yeah.

是的。

Yeah.

是的。

Yeah.

好的。

Okay.

对。

Yeah.

好吧。

Alright.

所以免疫系统，癌症，嗯。

So the immune system, the the cancer Mhmm.

这些是世界上主要的疾病成因。

These are leading causes of maladies in this world.

那些影响十万分之一人口的遗传病呢？

What of the genetic disorders that affect one in a hundred thousand people.

哇。

Wow.

你把它们加在一起，全球范围内患者数量足以填满一个体育场，但正因为太罕见，根本引不起人们的关注。

You you bring them together, there's enough of them, you know, they'll fill a stadium in the world, but that's so uncommon as to not really trigger anybody's interest.

这同时也是。

It's also yeah.

这也不赚钱。

It's also not profitable.

嗯，确实。

Well, yeah.

因为你没有足够的市场来销售这种药物。

Because because you don't have you don't have enough of a market there to sell the drug.

对。

Right.

我的意思是，查克，正是这一点。

I mean, know, it's exactly that that point, Chuck.

传统上，针对非常小的患者群体可能商业吸引力不大，你知道的。

It's, traditionally, it might not be that attractive commercially to go after, you know, very small patient populations.

但在一个获取这些药物分子变得便宜得多、容易得多的世界里，这就打开了所有这些可能性。

But in a world where it's so much cheaper, so much easier to get to these drug molecules, then that opens up, like, all of this space.

而且我们

And we we

越便宜越好，

like, the cheaper it is,

对。

the Right.

你就越能合理地去追求这条

The easier you can justify going down that

风险路径。

risk list.

而这正是

And this is

对我们来说的一个重要指引。

a big guiding star for us.

这就是我们这么做的原因。

This is why we're doing this.

这太美了。

That's beautiful.

我明白你的意思了。

I see what you did there.

指引之星。

Guiding star.

是的。

Yeah.

是的。

Yeah.

你知道，如果能

You know, it'd be great to

选择这些人和环境就好了。

pick these people the environment.

他并没有坐着。

He's not sitting.

首席AI官是怎么回事？

How Why is chief AI officer?

是的。

Yeah.

我对贵公司在塑造与您工作相关的政策方面的活跃程度很感兴趣，因为将会有大量立法政策与您的工作相关。

I'm interested to how active the company is in shaping policy around what you're doing because there's gonna be a great deal of legislative policy that is going to be tied to what you're doing.

所有的专利影响，还有研究方面的意义，都会随之而来。

All of the patent implications, there's gonna be, you know, research implications.

会有许多事情与此相关。

There's be a lot of things tied to this.

是的。

Yeah.

是的。

Yeah.

我的意思是，我们在这次对话中一直在讨论药物设计，但一旦设计出药物，就必须进入患者进行临床试验。

I mean, we we've been talking in this conversation about drug design, but then once you've designed the drug, you've gotta go into patients in clinical trials.

对。

Right.

这是一个非常漫长的过程。

And that's a really long process.

是的。

Yeah.

所以我们用小鼠做实验。

And That's why we hit mice.

但即使如此

But even

这些小鼠模型其实预测性并不强。

these mice models, they're not actually very predictive.

比如，你在小鼠身上做了所有这些研究，但结果往往无法在人体中重现。

Like, you do all these studies in mice, and then they don't you know, it doesn't translate into success in people.

对。

Right.

是的。

Yeah.

必须往上走，进入进化更高的层级，然后才能触及人类层面。

Gotta go up the evolutionary scale and then get to the human bit.

是的。

Yeah.

没错。

Exactly.

因此，你可以想象这样一个世界：我们可以设计出大量新药。

And so there's you know, you can imagine a world where we can design loads of new drugs.

我们必须改变临床试验的方式，想想如何才能及时地把这些药物送到真正需要它们的患者手中。

We've gotta be changing the way that we're doing clinical trials, you know, how we can actually get these drugs to patients who really, really need them in a timely manner.

所以我认为这方面还有很多工作要做，很多地方需要重新思考。

So I think there's a lot to be done and, like, rethought there.

AlphaFold以及整个医学科学的最终目标，是不是能够为每个人量身定制药物，而不是像现在这样使用针对广泛人群、却带来诸多副作用的通用药物？

Is the ultimate goal for AlphaFold and, I think, medical science as a whole to be able to bespoke medication for you as the individual rather than the broader spectrum medication that we find ourselves with all the side effects.

那么，你能否设计出一种对您完全有效且没有任何副作用的药物呢？

So are you able to then design a drug or a medication that has zero side effects and works exactly for me?

这就是目标。

This is the goal.

对吧？

Right?

这正是我们努力的方向。

This is this is what we're shooting for.

你可以想象这样一个世界：我们可以测序你特有的癌症突变。

You know, imagine a world where we can sequence your particular cancer mutations.

没错。

Right.

嗯。

Mhmm.

然后根据这些个体突变，为你生成专属的药物。

And then based on those, your individual mutations be generating specific drugs for you.

对。

Right.

就连这些药物也像是即将实现的三维打印技术一样。

That even these are like, you know, three d printed or something around the corner.

好的。

Okay.

是的。

Yeah.

我们现在正处于癌症免疫疗法的非常初期阶段，而且

This is We're in the very nascent stages of that right now with immunotherapy for cancer treatments and

但这些尚未被治愈的疾病中，有多少适合通过蛋白质折叠来解决？

But how many of these yet to be cured diseases lend themselves to solutions that involve protein folding?

又有多少只是传统的化学方法？

And how many are just plain old old fashioned chemistry?

蛋白质构成了我们几乎所有的分子机器。

Proteins make up, like, pretty much all of our molecular machinery.

对。

Right.

因此，有一类疾病是由于蛋白质错误折叠引起的，但还有很多其他疾病是由于蛋白质表达异常等原因造成的。

So there's there's a class of disease which is due to misfolding, but then there's many many other diseases which are due to, for example, a protein not being expressed properly

对。

Right.

或者，你知道的，某个细胞在某个方面出了问题。

Or, you know, a cell going wrong in a certain

我给自己用抗菌化学物质，然后就解决了。

tissue bacterial infection, I give myself antibacterial chemicals, and then I'm done.

那我还需要你吗？

Do I need you for that?

但这些化学物质是在与细菌中的蛋白质相互作用。

But those chemicals are interacting with the proteins in the

细菌。

bacteria.

好的。

Okay.

因此，蛋白质是基本的分子机器，而药物这类化学物质通过调节这些蛋白质发挥作用，无论是在我们的细胞中，还是在

So proteins are the fundamental machinery, and the chemicals which are drugs modulating those proteins, whether it's like in our cells, in

细菌中。

bacteria.

明白了。

Alright.

所以，你所谈论的一切基本上都是在细胞层面上发生的。

So it's basically everything you're talking about is is all happening on the cellular level.

是的。

Yes.

如果你所描述的过程发生在细胞内部，蛋白质在进行它们的三维拼图式运作，而你已经有了应对方案或疗法，那么你就必须让你的疗法进入细胞内部

If what you're describing is happening inside of cells, proteins doing their thing, their three d jigsaw puzzle, and you have a solution for that, a remedy, you have to get your remedy inside the cell

嗯。

Mhmm.

以与这种折叠相互作用。

To interact with that folding.

递送系统。

The delivery system.

除了像特洛伊木马病毒之类的方式，你还能怎么做到这一点？

And how do you do that other than through like a Trojan horse virus or something?

因为病毒可以很容易地进入细胞。

Because viruses get in there pretty on command.

是的。

Yeah.

好吧，如果你

Well, if you

想想你吃的药片，药物设计其实非常困难，因为这不仅仅是针对这些蛋白质。

think about the drugs that you take as pills, drug drug design is really hard because it's not just about targeting these proteins.

我们还得把它们送到正确的地方。

We gotta get them to the right place.