0:00:01.004,0:00:03.858
So how are we going to beat[br]this novel coronavirus?

0:00:04.091,0:00:06.950
By using our best tools:

0:00:07.217,0:00:09.585
our science and our technology.

0:00:09.802,0:00:12.989
In my lab, we're using[br]the tools of artificial intelligence

0:00:12.989,0:00:14.657
and synthetic biology

0:00:14.657,0:00:18.285
to speed up the fight[br]against this pandemic.

0:00:18.285,0:00:20.519
Our work was originally designed

0:00:20.519,0:00:23.089
to tackle the antibiotic[br]resistance crisis.

0:00:23.089,0:00:26.048
Our project seeks to harness[br]the power of machine learning

0:00:26.048,0:00:29.517
to replenish our antibiotic arsenal

0:00:29.517,0:00:33.120
and avoid a globally devastating[br]post-antibiotic era.

0:00:33.749,0:00:37.075
Importantly, the same technology[br]can be used to search

0:00:37.075,0:00:38.777
for antiviral compounds

0:00:38.777,0:00:41.512
that could help us fight[br]the current pandemic.

0:00:42.310,0:00:46.479
Machine learning is turning[br]the traditional model of drug discovery

0:00:46.479,0:00:47.529
on its head.

0:00:47.529,0:00:49.214
With this approach,

0:00:49.214,0:00:51.465
instead of painstakingly testing[br]thousands of existing molecules

0:00:51.465,0:00:53.217
one by one in a lab[br]for their effectiveness,

0:00:53.217,0:00:56.203
we can train a computer

0:00:56.203,0:01:00.430
to explore the exponentially larger space

0:01:00.430,0:01:04.435
of essentially all possible molecules[br]that could be synthesized,

0:01:04.435,0:01:10.014
and thus instead of looking[br]for a needle in a haystack,

0:01:10.014,0:01:13.783
we can use the giant magnet[br]of computing power

0:01:13.783,0:01:18.677
to find many needles[br]in multiple haystacks simultaneously.

0:01:18.677,0:01:20.462
We've already had some early success.

0:01:20.462,0:01:24.784
Recently, we used machine learning

0:01:24.784,0:01:28.076
to discover new antibiotics[br]that can help us fight off

0:01:28.076,0:01:32.547
the bacterial infections that can occur[br]alongside SARS-CoV-2 infections.

0:01:33.348,0:01:35.784
Two months ago, TED's Audacious Project

0:01:35.784,0:01:39.376
approved funding for us[br]to massively scale up our work

0:01:39.376,0:01:44.415
with the goal of discovering[br]seven new classes of antibiotics

0:01:44.415,0:01:47.384
against seven of the world's[br]deadly bacterial pathogens

0:01:47.992,0:01:50.528
over the next seven years.

0:01:50.731,0:01:52.428
For context,

0:01:52.428,0:01:55.071
the number of new class of antibiotics

0:01:55.071,0:01:58.290
that have been discovered[br]over the last three decades is zero.

0:01:58.404,0:02:01.927
While the quest for new antibiotics[br]is for our medium-term future,

0:02:01.927,0:02:06.435
the novel coronavirus[br]poses an immediate deadly threat,

0:02:06.602,0:02:10.388
and I'm excited to share that we think[br]we can use the same technology

0:02:10.388,0:02:13.548
to search for therapeutics[br]to fight this virus.

0:02:13.748,0:02:15.549
So how are we going to do it?

0:02:15.549,0:02:18.403
Well, we're creating[br]a compound training library,

0:02:18.403,0:02:21.644
and with collaborators[br]applying these molecules

0:02:21.644,0:02:23.957
to SARS-CoV-2-infected cells

0:02:23.957,0:02:28.398
to see which of them exhibit[br]effective activity.

0:02:28.398,0:02:31.538
These data will be use to train[br]a machine learning model

0:02:31.538,0:02:35.709
that will be applied to a [?][br]library of over a billion molecules

0:02:35.709,0:02:40.570
to search for potential[br]novel antiviral compounds.

0:02:40.570,0:02:43.289
We will synthesize and test[br]the top predictions

0:02:43.289,0:02:46.275
and advance the most promising[br]candidates into the clinic.

0:02:46.508,0:02:48.701
Sound too good to be true?

0:02:48.701,0:02:49.869
Well, it shouldn't.

0:02:49.869,0:02:53.276
The Antibiotics AI Project is founded[br]on our proof of concept research

0:02:53.276,0:02:56.674
that led to the discovery[br]of a novel broad spectrum antibiotic

0:02:56.674,0:02:57.859
called Halocin.

0:02:58.642,0:03:01.478
Halocin has potent antibacterial activity

0:03:01.478,0:03:05.615
against almost all antibiotic-resistant[br]bacterial pathogens,

0:03:05.801,0:03:09.442
including untreatable[br]pan-resistant infections.

0:03:09.942,0:03:12.211
Importantly, in contrast[br]to current antibiotics,

0:03:12.211,0:03:16.215
the frequency at which bacteria[br]develop resistance against Halocin

0:03:16.215,0:03:17.673
is remarkably low.

0:03:18.224,0:03:23.178
We tested the ability of bacteria[br]to evolve resistance against Halocin

0:03:23.178,0:03:25.299
as well as Cipro in the lab.

0:03:25.299,0:03:27.131
In the case of Cipro,

0:03:27.131,0:03:30.175
after just one day, we saw resistance.

0:03:30.391,0:03:31.960
In the case of Halocin,

0:03:31.960,0:03:34.646
after one day we didn't[br]see any resistance.

0:03:34.646,0:03:38.039
Amazingly, after even 30 days,

0:03:38.039,0:03:40.991
we didn't see any[br]resistance against Halocin.

0:03:41.242,0:03:46.980
In this pilot project, we first tested[br]roughly 2,500 compounds against E. coli.

0:03:47.213,0:03:50.273
This training set included[br]known antibiotics,

0:03:50.273,0:03:52.176
such as Cipro and penicillin,

0:03:52.176,0:03:55.127
as well as many drugs[br]that are not antibiotics.

0:03:55.294,0:03:57.904
These data we used to train a model

0:03:57.904,0:04:02.491
to learn molecular features[br]associated with antibacterial activity.

0:04:02.491,0:04:05.542
We then applied this model[br]to a drug repurposing library

0:04:05.995,0:04:08.187
consisting of several thousand molecules,

0:04:08.187,0:04:10.456
and asked the model to identify molecules

0:04:10.456,0:04:13.225
that are predicted[br]to have antibacterial properties

0:04:13.225,0:04:15.910
but don't look like existing antibiotics.

0:04:16.194,0:04:21.305
Interestingly, only one molecule[br]in that library fit these criteria,

0:04:21.506,0:04:24.342
and that molecule[br]turned out to be Halocin.

0:04:24.691,0:04:27.769
Given that Halocin does not look[br]like any existing antibiotic,

0:04:27.769,0:04:31.822
it would have been impossible for a human,[br]including an antibiotic expert,

0:04:31.822,0:04:34.724
to identify Halocin in this manner.

0:04:34.892,0:04:37.545
Imagine now what we could do[br]with this technology

0:04:37.545,0:04:40.053
against SARS-CoV-2.

0:04:40.053,0:04:43.106
And that's not all.

0:04:43.106,0:04:44.423
We're also using the tools[br]of synthetic biology,

0:04:44.423,0:04:46.893
tinkering with DNA[br]and other cellular machinery,

0:04:46.893,0:04:50.051
to serve human purposes[br]like combating COVID-19,

0:04:50.402,0:04:54.312
and at [??] we are working[br]to develop a protective mask

0:04:54.411,0:04:58.193
that can also serve[br]as a rapid diagnostic test.

0:04:58.414,0:04:59.945
So how does that work?

0:04:59.945,0:05:02.617
Well, we recently showed that you can take

0:05:02.617,0:05:04.360
the cellular machinery[br]out of a living cell

0:05:04.360,0:05:07.912
and freeze-dry it along with[br]RNA sensors onto paper

0:05:08.261,0:05:12.916
in order to create low-cost[br]diagnostics for Ebola and Zika.

0:05:13.750,0:05:18.429
The sensors are activated when[br]they're rehydrated by a patient sample

0:05:18.429,0:05:21.563
that could consist of blood[br]or saliva, for example.

0:05:21.781,0:05:25.040
It turns out this technology[br]is not limited to paper

0:05:25.040,0:05:28.894
and can be applied[br]to other materials, including cloth.

0:05:28.894,0:05:31.163
For the COVID-19 pandemic,[br]we're designing RNA sensors

0:05:31.163,0:05:33.014
to detect the virus

0:05:33.014,0:05:38.412
and freeze-drying these[br]along with the needed cellular machinery

0:05:38.412,0:05:40.561
into the fabric of a face mask,

0:05:40.561,0:05:42.851
where the simple act of breathing,

0:05:43.414,0:05:47.725
along with the water vapor[br]that comes with it, can activate the test.

0:05:48.043,0:05:51.496
Now, if the patient[br]is infected with SARS-CoV-2,

0:05:52.263,0:05:56.350
the mask will produce a fluorescent signal[br]that can be detected by a simple,

0:05:56.350,0:05:58.860
inexpensive handheld device.

0:05:58.860,0:06:01.111
In one or two hours, a patient[br]could those be diagnosed

0:06:01.111,0:06:06.339
safely, remotely and accurately.

0:06:07.006,0:06:10.349
We're also using synthetic biology

0:06:10.349,0:06:12.678
to design a candidate[br]vaccine for COVID-19.

0:06:12.912,0:06:15.830
We are repurposing the BCG vaccine,

0:06:15.830,0:06:18.975
which has been used against TB[br]for almost a century.

0:06:18.975,0:06:20.459
It's a live attenuated vaccine,

0:06:20.459,0:06:24.228
and we're engineering it[br]to express SARS-CoV-2 antigens,

0:06:24.777,0:06:27.829
which should trigger the production[br]of protective antibodies

0:06:27.829,0:06:29.649
by the immune system.

0:06:29.649,0:06:33.544
Importantly, BCG is massively scalable[br]and has a safety profile

0:06:33.544,0:06:37.012
that's among the best[br]of any reported vaccine.

0:06:38.096,0:06:43.224
With the tools of synthetic biology[br]and artificial intelligence,

0:06:43.224,0:06:46.611
we can win the fight[br]against this novel coronavirus.

0:06:46.844,0:06:50.354
This work is in its very early stages,[br]but the promise is real.

0:06:50.354,0:06:54.591
Science and technology[br]can give us an important advantage

0:06:54.591,0:06:57.695
in the battle of human wits[br]versus the genes of superbugs,

0:06:57.695,0:06:59.871
a battle we can win.

0:06:59.871,0:07:01.655
Thank you.