[Script Info]
Title: 
[Events]
Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
Dialogue: 0,0:00:00.92,0:00:03.82,Default,,0000,0000,0000,,So how are we going to beat\Nthis novel coronavirus?
Dialogue: 0,0:00:04.32,0:00:06.95,Default,,0000,0000,0000,,By using our best tools:
Dialogue: 0,0:00:06.97,0:00:09.01,Default,,0000,0000,0000,,our science and our technology.
Dialogue: 0,0:00:09.59,0:00:12.73,Default,,0000,0000,0000,,In my lab, we're using\Nthe tools of artificial intelligence
Dialogue: 0,0:00:12.75,0:00:14.33,Default,,0000,0000,0000,,and synthetic biology
Dialogue: 0,0:00:14.35,0:00:17.41,Default,,0000,0000,0000,,to speed up the fight\Nagainst this pandemic.
Dialogue: 0,0:00:18.08,0:00:19.94,Default,,0000,0000,0000,,Our work was originally designed
Dialogue: 0,0:00:19.96,0:00:22.82,Default,,0000,0000,0000,,to tackle the antibiotic\Nresistance crisis.
Dialogue: 0,0:00:22.84,0:00:27.53,Default,,0000,0000,0000,,Our project seeks to harness\Nthe power of machine learning
Dialogue: 0,0:00:27.56,0:00:29.40,Default,,0000,0000,0000,,to replenish our antibiotic arsenal
Dialogue: 0,0:00:29.42,0:00:33.26,Default,,0000,0000,0000,,and avoid a globally devastating\Npostantibiotic era.
Dialogue: 0,0:00:33.68,0:00:36.50,Default,,0000,0000,0000,,Importantly, the same\Ntechnology can be used
Dialogue: 0,0:00:36.53,0:00:38.60,Default,,0000,0000,0000,,to search for antiviral compounds
Dialogue: 0,0:00:38.62,0:00:41.30,Default,,0000,0000,0000,,that could help us fight\Nthe current pandemic.
Dialogue: 0,0:00:42.08,0:00:45.98,Default,,0000,0000,0000,,Machine learning is turning\Nthe traditional model of drug discovery
Dialogue: 0,0:00:46.01,0:00:47.41,Default,,0000,0000,0000,,on its head.
Dialogue: 0,0:00:47.43,0:00:48.66,Default,,0000,0000,0000,,With this approach,
Dialogue: 0,0:00:48.68,0:00:52.76,Default,,0000,0000,0000,,instead of painstakingly testing\Nthousands of existing molecules
Dialogue: 0,0:00:52.78,0:00:54.22,Default,,0000,0000,0000,,one by one in a lab
Dialogue: 0,0:00:54.24,0:00:55.83,Default,,0000,0000,0000,,for their effectiveness,
Dialogue: 0,0:00:55.86,0:01:00.51,Default,,0000,0000,0000,,we can train a computer\Nto explore the exponentially larger space
Dialogue: 0,0:01:00.54,0:01:04.12,Default,,0000,0000,0000,,of essentially all possible molecules\Nthat could be synthesized,
Dialogue: 0,0:01:04.14,0:01:09.76,Default,,0000,0000,0000,,and thus, instead of looking\Nfor a needle in a haystack,
Dialogue: 0,0:01:09.78,0:01:13.54,Default,,0000,0000,0000,,we can use the giant magnet\Nof computing power
Dialogue: 0,0:01:13.57,0:01:17.48,Default,,0000,0000,0000,,to find many needles\Nin multiple haystacks simultaneously.
Dialogue: 0,0:01:18.42,0:01:20.42,Default,,0000,0000,0000,,We've already had some early success.
Dialogue: 0,0:01:21.01,0:01:26.48,Default,,0000,0000,0000,,Recently, we used machine learning\Nto discover new antibiotics
Dialogue: 0,0:01:26.50,0:01:29.06,Default,,0000,0000,0000,,that can help us fight off\Nthe bacterial infections
Dialogue: 0,0:01:29.08,0:01:32.69,Default,,0000,0000,0000,,that can occur alongside\NSARS-CoV-2 infections.
Dialogue: 0,0:01:33.18,0:01:37.35,Default,,0000,0000,0000,,Two months ago, TED's Audacious Project\Napproved funding for us
Dialogue: 0,0:01:37.37,0:01:39.56,Default,,0000,0000,0000,,to massively scale up our work
Dialogue: 0,0:01:39.59,0:01:44.21,Default,,0000,0000,0000,,with the goal of discovering\Nseven new classes of antibiotics
Dialogue: 0,0:01:44.24,0:01:47.72,Default,,0000,0000,0000,,against seven of the world's\Ndeadly bacterial pathogens
Dialogue: 0,0:01:47.74,0:01:49.80,Default,,0000,0000,0000,,over the next seven years.
Dialogue: 0,0:01:50.21,0:01:51.94,Default,,0000,0000,0000,,For context:
Dialogue: 0,0:01:51.96,0:01:53.89,Default,,0000,0000,0000,,the number of new class of antibiotics
Dialogue: 0,0:01:53.92,0:01:57.15,Default,,0000,0000,0000,,that have been discovered\Nover the last three decades is zero.
Dialogue: 0,0:01:58.03,0:02:01.60,Default,,0000,0000,0000,,While the quest for new antibiotics\Nis for our medium-term future,
Dialogue: 0,0:02:01.62,0:02:06.28,Default,,0000,0000,0000,,the novel coronavirus poses\Nan immediate deadly threat,
Dialogue: 0,0:02:06.30,0:02:10.09,Default,,0000,0000,0000,,and I'm excited to share that we think\Nwe can use the same technology
Dialogue: 0,0:02:10.12,0:02:12.93,Default,,0000,0000,0000,,to search for therapeutics\Nto fight this virus.
Dialogue: 0,0:02:13.49,0:02:15.20,Default,,0000,0000,0000,,So how are we going to do it?
Dialogue: 0,0:02:15.23,0:02:18.18,Default,,0000,0000,0000,,Well, we're creating\Na compound training library
Dialogue: 0,0:02:18.20,0:02:23.74,Default,,0000,0000,0000,,and with collaborators applying these\Nmolecules to SARS-CoV-2-infected cells
Dialogue: 0,0:02:23.77,0:02:27.66,Default,,0000,0000,0000,,to see which of them exhibit\Neffective activity.
Dialogue: 0,0:02:28.18,0:02:31.37,Default,,0000,0000,0000,,These data will be use to train\Na machine learning model
Dialogue: 0,0:02:31.39,0:02:35.46,Default,,0000,0000,0000,,that will be applied to an in silico\Nlibrary of over a billion molecules
Dialogue: 0,0:02:35.48,0:02:39.69,Default,,0000,0000,0000,,to search for potential\Nnovel antiviral compounds.
Dialogue: 0,0:02:40.32,0:02:42.98,Default,,0000,0000,0000,,We will synthesize and test\Nthe top predictions
Dialogue: 0,0:02:43.01,0:02:45.90,Default,,0000,0000,0000,,and advance the most promising\Ncandidates into the clinic.
Dialogue: 0,0:02:46.36,0:02:48.13,Default,,0000,0000,0000,,Sound too good to be true?
Dialogue: 0,0:02:48.16,0:02:49.59,Default,,0000,0000,0000,,Well, it shouldn't.
Dialogue: 0,0:02:49.61,0:02:52.94,Default,,0000,0000,0000,,The Antibiotics AI Project is founded\Non our proof of concept research
Dialogue: 0,0:02:52.96,0:02:56.36,Default,,0000,0000,0000,,that led to the discovery\Nof a novel broad-spectrum antibiotic
Dialogue: 0,0:02:56.39,0:02:57.57,Default,,0000,0000,0000,,called halicin.
Dialogue: 0,0:02:58.44,0:03:01.26,Default,,0000,0000,0000,,Halicin has potent antibacterial activity
Dialogue: 0,0:03:01.28,0:03:05.38,Default,,0000,0000,0000,,against almost all antibiotic-resistant\Nbacterial pathogens,
Dialogue: 0,0:03:05.41,0:03:09.05,Default,,0000,0000,0000,,including untreatable\Npanresistant infections.
Dialogue: 0,0:03:09.86,0:03:12.13,Default,,0000,0000,0000,,Importantly, in contrast\Nto current antibiotics,
Dialogue: 0,0:03:12.16,0:03:15.85,Default,,0000,0000,0000,,the frequency at which bacteria\Ndevelop resistance against halicin
Dialogue: 0,0:03:15.87,0:03:17.36,Default,,0000,0000,0000,,is remarkably low.
Dialogue: 0,0:03:18.30,0:03:23.01,Default,,0000,0000,0000,,We tested the ability of bacteria\Nto evolve resistance against halicin
Dialogue: 0,0:03:23.04,0:03:24.82,Default,,0000,0000,0000,,as well as Cipro in the lab.
Dialogue: 0,0:03:25.30,0:03:26.84,Default,,0000,0000,0000,,In the case of Cipro,
Dialogue: 0,0:03:26.86,0:03:29.69,Default,,0000,0000,0000,,after just one day, we saw resistance.
Dialogue: 0,0:03:30.21,0:03:31.69,Default,,0000,0000,0000,,In the case of halicin,
Dialogue: 0,0:03:31.72,0:03:33.83,Default,,0000,0000,0000,,after one day,\Nwe didn't see any resistance.
Dialogue: 0,0:03:34.48,0:03:37.78,Default,,0000,0000,0000,,Amazingly, after even 30 days,
Dialogue: 0,0:03:37.80,0:03:40.41,Default,,0000,0000,0000,,we didn't see any resistance\Nagainst halicin.
Dialogue: 0,0:03:41.10,0:03:46.62,Default,,0000,0000,0000,,In this pilot project, we first tested\Nroughly 2,500 compounds against E. coli.
Dialogue: 0,0:03:47.26,0:03:50.04,Default,,0000,0000,0000,,This training set included\Nknown antibiotics,
Dialogue: 0,0:03:50.06,0:03:51.81,Default,,0000,0000,0000,,such as Cipro and penicillin,
Dialogue: 0,0:03:51.83,0:03:54.10,Default,,0000,0000,0000,,as well as many drugs\Nthat are not antibiotics.
Dialogue: 0,0:03:54.98,0:03:57.57,Default,,0000,0000,0000,,These data we used to train a model
Dialogue: 0,0:03:57.60,0:04:01.57,Default,,0000,0000,0000,,to learn molecular features\Nassociated with antibacterial activity.
Dialogue: 0,0:04:02.27,0:04:04.97,Default,,0000,0000,0000,,We then applied this model\Nto a drug-repurposing library
Dialogue: 0,0:04:04.99,0:04:07.47,Default,,0000,0000,0000,,consisting of several thousand molecules
Dialogue: 0,0:04:07.50,0:04:10.11,Default,,0000,0000,0000,,and asked the model to identify molecules
Dialogue: 0,0:04:10.14,0:04:12.92,Default,,0000,0000,0000,,that are predicted\Nto have antibacterial properties
Dialogue: 0,0:04:12.95,0:04:15.42,Default,,0000,0000,0000,,but don't look like existing antibiotics.
Dialogue: 0,0:04:16.43,0:04:21.22,Default,,0000,0000,0000,,Interestingly, only one molecule\Nin that library fit these criteria,
Dialogue: 0,0:04:21.25,0:04:23.58,Default,,0000,0000,0000,,and that molecule\Nturned out to be halicin.
Dialogue: 0,0:04:24.44,0:04:27.53,Default,,0000,0000,0000,,Given that halicin does not look\Nlike any existing antibiotic,
Dialogue: 0,0:04:27.56,0:04:31.71,Default,,0000,0000,0000,,it would have been impossible for a human,\Nincluding an antibiotic expert,
Dialogue: 0,0:04:31.73,0:04:33.92,Default,,0000,0000,0000,,to identify halicin in this manner.
Dialogue: 0,0:04:34.57,0:04:37.20,Default,,0000,0000,0000,,Imagine now what we could do\Nwith this technology
Dialogue: 0,0:04:37.23,0:04:38.97,Default,,0000,0000,0000,,against SARS-CoV-2.
Dialogue: 0,0:04:39.78,0:04:41.15,Default,,0000,0000,0000,,And that's not all.
Dialogue: 0,0:04:41.17,0:04:43.99,Default,,0000,0000,0000,,We're also using the tools\Nof synthetic biology,
Dialogue: 0,0:04:44.02,0:04:46.63,Default,,0000,0000,0000,,tinkering with DNA\Nand other cellular machinery,
Dialogue: 0,0:04:46.65,0:04:50.56,Default,,0000,0000,0000,,to serve human purposes\Nlike combating COVID-19,
Dialogue: 0,0:04:50.58,0:04:54.23,Default,,0000,0000,0000,,and of note, we are working\Nto develop a protective mask
Dialogue: 0,0:04:54.26,0:04:57.69,Default,,0000,0000,0000,,that can also serve\Nas a rapid diagnostic test.
Dialogue: 0,0:04:58.19,0:04:59.66,Default,,0000,0000,0000,,So how does that work?
Dialogue: 0,0:04:59.69,0:05:00.89,Default,,0000,0000,0000,,Well, we recently showed
Dialogue: 0,0:05:00.92,0:05:03.86,Default,,0000,0000,0000,,that you can take the cellular\Nmachinery out of a living cell
Dialogue: 0,0:05:03.88,0:05:07.98,Default,,0000,0000,0000,,and freeze-dry it along with\NRNA sensors onto paper
Dialogue: 0,0:05:08.00,0:05:12.92,Default,,0000,0000,0000,,in order to create low-cost\Ndiagnostics for Ebola and Zika.
Dialogue: 0,0:05:13.50,0:05:18.73,Default,,0000,0000,0000,,The sensors are activated when\Nthey're rehydrated by a patient sample
Dialogue: 0,0:05:18.75,0:05:21.58,Default,,0000,0000,0000,,that could consist of blood\Nor saliva, for example.
Dialogue: 0,0:05:21.60,0:05:24.86,Default,,0000,0000,0000,,It turns out, this technology\Nis not limited to paper
Dialogue: 0,0:05:24.88,0:05:27.77,Default,,0000,0000,0000,,and can be applied\Nto other materials, including cloth.
Dialogue: 0,0:05:28.67,0:05:30.61,Default,,0000,0000,0000,,For the COVID-19 pandemic,
Dialogue: 0,0:05:30.64,0:05:34.98,Default,,0000,0000,0000,,we're designing RNA sensors\Nto detect the virus
Dialogue: 0,0:05:35.01,0:05:38.22,Default,,0000,0000,0000,,and freeze-drying these\Nalong with the needed cellular machinery
Dialogue: 0,0:05:38.24,0:05:40.95,Default,,0000,0000,0000,,into the fabric of a face mask,
Dialogue: 0,0:05:40.97,0:05:43.20,Default,,0000,0000,0000,,where the simple act of breathing,
Dialogue: 0,0:05:43.22,0:05:45.50,Default,,0000,0000,0000,,along with the water vapor\Nthat comes with it,
Dialogue: 0,0:05:45.53,0:05:47.29,Default,,0000,0000,0000,,can activate the test.
Dialogue: 0,0:05:47.80,0:05:52.06,Default,,0000,0000,0000,,Thus, if a patient is infected\Nwith SARS-CoV-2,
Dialogue: 0,0:05:52.09,0:05:54.16,Default,,0000,0000,0000,,the mask will produce\Na fluorescent signal
Dialogue: 0,0:05:54.18,0:05:58.02,Default,,0000,0000,0000,,that could be detected by a simple,\Ninexpensive handheld device.
Dialogue: 0,0:05:58.53,0:06:03.02,Default,,0000,0000,0000,,In one or two hours, a patient\Ncould thus be diagnosed
Dialogue: 0,0:06:03.04,0:06:06.01,Default,,0000,0000,0000,,safely, remotely and accurately.
Dialogue: 0,0:06:06.74,0:06:09.26,Default,,0000,0000,0000,,We're also using synthetic biology
Dialogue: 0,0:06:09.28,0:06:11.100,Default,,0000,0000,0000,,to design a candidate\Nvaccine for COVID-19.
Dialogue: 0,0:06:13.01,0:06:15.67,Default,,0000,0000,0000,,We are repurposing the BCG vaccine,
Dialogue: 0,0:06:15.69,0:06:18.56,Default,,0000,0000,0000,,which had been used against TB\Nfor almost a century.
Dialogue: 0,0:06:18.58,0:06:20.13,Default,,0000,0000,0000,,It's a live attenuated vaccine,
Dialogue: 0,0:06:20.15,0:06:24.81,Default,,0000,0000,0000,,and we're engineering it\Nto express SARS-CoV-2 antigens,
Dialogue: 0,0:06:24.83,0:06:27.64,Default,,0000,0000,0000,,which should trigger the production\Nof protective antibodies
Dialogue: 0,0:06:27.67,0:06:29.30,Default,,0000,0000,0000,,by the immune system.
Dialogue: 0,0:06:29.33,0:06:32.06,Default,,0000,0000,0000,,Importantly, BCG\Nis massively scalable
Dialogue: 0,0:06:32.09,0:06:36.66,Default,,0000,0000,0000,,and has a safety profile that's among\Nthe best of any reported vaccine.
Dialogue: 0,0:06:37.88,0:06:42.99,Default,,0000,0000,0000,,With the tools of synthetic biology\Nand artificial intelligence,
Dialogue: 0,0:06:43.01,0:06:46.36,Default,,0000,0000,0000,,we can win the fight\Nagainst this novel coronavirus.
Dialogue: 0,0:06:46.84,0:06:50.16,Default,,0000,0000,0000,,This work is in its very early stages,\Nbut the promise is real.
Dialogue: 0,0:06:50.80,0:06:54.24,Default,,0000,0000,0000,,Science and technology\Ncan give us an important advantage
Dialogue: 0,0:06:54.27,0:06:57.43,Default,,0000,0000,0000,,in the battle of human wits\Nversus the genes of superbugs,
Dialogue: 0,0:06:57.45,0:06:59.20,Default,,0000,0000,0000,,a battle we can win.
Dialogue: 0,0:06:59.99,0:07:01.22,Default,,0000,0000,0000,,Thank you.