How we're using AI to discover new antibiotics | Jim Collins

翻译人员: Jinhao Ma 校对人员: Wanting Zhong

So how are we going to beat this novel coronavirus?

我们要如何击败新型冠状病毒？

By using our best tools:

通过使用我们最好的工具：

our science and our technology.

我们的科学和技术。

In my lab, we're using the tools of artificial intelligence

在我的实验室中， 我们正在使用人工智能

and synthetic biology

和合成生物学的工具，

to speed up the fight against this pandemic.

加快与这场疫情的战斗。

Our work was originally designed

我们工作的初衷

to tackle the antibiotic resistance crisis.

是想解决抗生素耐药性的危机。

Our project seeks to harness the power of machine learning

我们的项目试图利用 机器学习的力量

to replenish our antibiotic arsenal

补充我们的抗生素“弹药库”，

and avoid a globally devastating postantibiotic era.

并避免会造成全球性危害的 后抗生素时代。

Importantly, the same technology can be used

重要的是，同样的技术能用来寻找

to search for antiviral compounds

可以帮助我们应对当前疫情的

that could help us fight the current pandemic.

抗病毒化合物。

Machine learning is turning the traditional model of drug discovery

机器学习正在颠覆

传统的药物开发模型。

on its head.

With this approach,

通过这种方法，

instead of painstakingly testing thousands of existing molecules

我们不再需要在实验室里 一个接一个费力地测试

one by one in a lab

成千上万

现有分子的效力，

for their effectiveness,

we can train a computer to explore the exponentially larger space

而是可以训练电脑探索更大的、

of essentially all possible molecules that could be synthesized,

基本上涵盖了所有 可能合成的分子的空间。

and thus, instead of looking for a needle in a haystack,

因此，相比在“海底捞针”，

we can use the giant magnet of computing power

我们可以使用计算能力 这块“巨型磁铁”，

to find many needles in multiple haystacks simultaneously.

同时在几个“海”底 捞很多很多根“针”。

We've already had some early success.

我们的早期尝试 已经取得了一些成功。

Recently, we used machine learning to discover new antibiotics

最近，我们使用机器学习 发现了新的抗生素，

that can help us fight off the bacterial infections

可以帮助我们抵御

可能伴随 SARS-CoV-2 冠状病毒感染 发生的细菌感染。

that can occur alongside SARS-CoV-2 infections.

Two months ago, TED's Audacious Project approved funding for us

两个月前，TED 的“大胆计划” （Audacious Project）

to massively scale up our work

批准了我们的资金申请，

with the goal of discovering seven new classes of antibiotics

这将大规模扩展我们的工作， 目标是在未来的七年里，

against seven of the world's deadly bacterial pathogens

发现七类新型抗生素，

以对抗世界上七种 致命的病原体细菌。

over the next seven years.

For context:

在此说明一下：

the number of new class of antibiotics

在过去三十年内，人类发现的

that have been discovered over the last three decades is zero.

新型抗生素的数量为零。

While the quest for new antibiotics is for our medium-term future,

虽说寻找新的抗生素 是为了我们的中期未来，

the novel coronavirus poses an immediate deadly threat,

新型冠状病毒构成了 迫在眉睫的致命威胁，

and I'm excited to share that we think we can use the same technology

我很高兴能跟大家宣布， 我们认为可以使用相同的技术

to search for therapeutics to fight this virus.

寻找对抗这种病毒的治疗手段。

So how are we going to do it?

那么我们该怎么做呢？

Well, we're creating a compound training library

我们正在创建一个 化合物训练库，

and with collaborators applying these molecules to SARS-CoV-2-infected cells

并与合作者一起，用这些分子处理 被 SARS-CoV-2 感染的细胞，

to see which of them exhibit effective activity.

看看哪个分子表现出了有效的活性。

These data will be use to train a machine learning model

这些数据将用于训练 一个机器学习模型，

that will be applied to an in silico library of over a billion molecules

这个模型将被应用于包含 超过十亿个分子的计算机模拟数据库，

to search for potential novel antiviral compounds.

以寻找潜在的新型抗病毒化合物。

We will synthesize and test the top predictions

我们将合成并测试 算法预测出的最优分子，

and advance the most promising candidates into the clinic.

并让最有潜力的备选分子 进入临床实验。

Sound too good to be true?

听起来是不是过于美好了？

Well, it shouldn't.

并非如此。

The Antibiotics AI Project is founded on our proof of concept research

抗生素人工智能项目的设立 是基于我们的概念验证研究，

that led to the discovery of a novel broad-spectrum antibiotic

这项研究最终发现了 一种新型广谱抗生素，

called halicin.

叫做 Halocin。

Halicin has potent antibacterial activity

Halocin 具有强大的抗菌活性，

against almost all antibiotic-resistant bacterial pathogens,

能杀死几乎所有 对抗生素耐药的病原体细菌，

包括无法治疗的多重耐药感染。

including untreatable panresistant infections.

Importantly, in contrast to current antibiotics,

重要的是，与目前的抗生素相比，

细菌对 Halocin 产生耐药性的频率

the frequency at which bacteria develop resistance against halicin

非常低。

is remarkably low.

We tested the ability of bacteria to evolve resistance against halicin

我们在实验室里测试了 细菌对 Halocin

以及环丙沙星（Cipro） 产生耐药性的能力。

as well as Cipro in the lab.

In the case of Cipro,

结果发现，

after just one day, we saw resistance.

仅仅一天后，细菌就对 环丙沙星产生了耐药性。

In the case of halicin,

而对于 Halocin，

after one day, we didn't see any resistance.

经过一天后， 细菌没有产生任何耐药性。

Amazingly, after even 30 days,

不可思议的是， 甚至在 30 天后，

we didn't see any resistance against halicin.

我们也没有发现细菌 对 Halocin 产生任何耐药性。

在这个试点项目中，我们首先对大肠杆菌 测试了大约 2500 种化合物。

In this pilot project, we first tested roughly 2,500 compounds against E. coli.

This training set included known antibiotics,

这个训练集包括了已知的抗生素，

例如环丙沙星和青霉素，

such as Cipro and penicillin,

as well as many drugs that are not antibiotics.

以及许多不是抗生素的药物。

These data we used to train a model

我们用这些数据来训练模型，

to learn molecular features associated with antibacterial activity.

让它学习与抗菌活性 有关的分子特征。

然后我们把这个模型 应用到由数千个分子组成的

We then applied this model to a drug-repurposing library

consisting of several thousand molecules

药物再定位数据库上，

and asked the model to identify molecules

并要求模型识别

被预测具有抗菌性能

that are predicted to have antibacterial properties

but don't look like existing antibiotics.

但长得不像现有抗生素的分子。

Interestingly, only one molecule in that library fit these criteria,

有趣的是，数据库里 只有一个分子符合这些条件，

and that molecule turned out to be halicin.

那个分子就是 Halocin。

Given that halicin does not look like any existing antibiotic,

由于 Halocin 看起来 不像任何现有的抗生素，

it would have been impossible for a human, including an antibiotic expert,

人类，包括抗生素专家，

to identify halicin in this manner.

都不可能以这种方式 发现 Halocin 的。

Imagine now what we could do with this technology

想象一下，我们能如何使用这项技术

against SARS-CoV-2.

对抗 SARS-CoV-2。

And that's not all.

还不止这些。

We're also using the tools of synthetic biology,

我们也在使用合成生物学的工具

修补 DNA 和其他细胞成分，

tinkering with DNA and other cellular machinery,

to serve human purposes like combating COVID-19,

为人类服务，比如对抗 COVID-19。

and of note, we are working to develop a protective mask

值得一提的是，我们正在努力开发

that can also serve as a rapid diagnostic test.

可作为快速诊断测试的防护口罩。

So how does that work?

它的原理是什么？

Well, we recently showed

我们最近发现

that you can take the cellular machinery out of a living cell

你可以从活细胞中 提取出细胞成分，

and freeze-dry it along with RNA sensors onto paper

然后把它连同 RNA 检测器 在试纸上进行冷冻干燥，

in order to create low-cost diagnostics for Ebola and Zika.

从而制作出廉价的 埃博拉和寨卡病毒诊断测试工具。

The sensors are activated when they're rehydrated by a patient sample

在通过添加患者的样本， 如血液或唾液进行重新溶解后，

that could consist of blood or saliva, for example.

RNA 检测器就能被激活。

It turns out, this technology is not limited to paper

事实证明，除了纸制品，

and can be applied to other materials, including cloth.

这项技术还可以应用于 其他材料，包括布料。

For the COVID-19 pandemic,

对于 COVID-19 疫情，

we're designing RNA sensors to detect the virus

我们正在设计 针对病毒的 RNA 检测器，

然后把它们和所需的细胞成分一起

and freeze-drying these along with the needed cellular machinery

into the fabric of a face mask,

在口罩的面料上进行冷冻干燥，

where the simple act of breathing,

简单的呼吸行为

along with the water vapor that comes with it,

连同呼出的水蒸气，

can activate the test.

就可以激活测试。

Thus, if a patient is infected with SARS-CoV-2,

如果患者感染了 SARS-CoV-2，

口罩就会产生荧光信号，

the mask will produce a fluorescent signal

that could be detected by a simple, inexpensive handheld device.

可以通过简单廉价的 手持设备检测出来。

In one or two hours, a patient could thus be diagnosed

一两个小时内，病人就能得到

安全、准确、无接触的诊断。

safely, remotely and accurately.

We're also using synthetic biology

我们也在使用合成生物学

to design a candidate vaccine for COVID-19.

设计 COVID-19 的备选疫苗。

We are repurposing the BCG vaccine,

我们正在重新利用卡介苗，

which had been used against TB for almost a century.

这种疫苗在近一个世纪前 就被用来预防结核病。

It's a live attenuated vaccine,

这是一种减毒活疫苗，

and we're engineering it to express SARS-CoV-2 antigens,

我们通过生物工程 让它表达 SARS-CoV-2 抗原，

which should trigger the production of protective antibodies

以此来触发免疫系统

by the immune system.

产生保护性抗体。

Importantly, BCG is massively scalable

重要的是，卡介苗可大规模生产，

并且它的安全性在所有 有记录的疫苗中是最好的。

and has a safety profile that's among the best of any reported vaccine.

With the tools of synthetic biology and artificial intelligence,

借助合成生物学与人工智能的工具，

we can win the fight against this novel coronavirus.

我们可以打赢 和新型冠状病毒的战争。

This work is in its very early stages, but the promise is real.

这项工作尚处于初期阶段， 但它的前景是真实的。

Science and technology can give us an important advantage

在人类智慧与超级细菌基因的战斗中，

in the battle of human wits versus the genes of superbugs,

科学和技术能给予我们重要的优势，

a battle we can win.

帮助我们取得胜利。

Thank you.

谢谢。