0:00:01.992,0:00:03.317
You don't know them.

0:00:04.150,0:00:05.436
You don't see them.

0:00:06.405,0:00:08.408
But they're always around,

0:00:09.404,0:00:11.084
whispering,

0:00:11.108,0:00:12.922
making secret plans,

0:00:13.700,0:00:17.382
building armies with millions of soldiers.

0:00:18.826,0:00:20.617
And when they decide to attack,

0:00:21.435,0:00:23.697
they all attack at the same time.

0:00:27.130,0:00:28.909
I'm talking about bacteria.

0:00:28.933,0:00:30.258
(Laughter)

0:00:30.282,0:00:32.137
Who did you think I was talking about?

0:00:34.401,0:00:37.595
Bacteria live in communities[br]just like humans.

0:00:37.619,0:00:38.892
They have families,

0:00:38.916,0:00:40.067
they talk,

0:00:40.091,0:00:41.933
and they plan their activities.

0:00:41.957,0:00:44.614
And just like humans, they trick, deceive,

0:00:44.638,0:00:46.772
and some might even cheat on each other.

0:00:48.127,0:00:51.974
What if I tell you that we can listen[br]to bacterial conversations

0:00:51.998,0:00:55.572
and translate their confidential[br]information into human language?

0:00:56.255,0:01:01.053
And what if I tell you that translating[br]bacterial conversations can save lives?

0:01:02.519,0:01:04.290
I hold a PhD in nanophysics,

0:01:04.314,0:01:08.690
and I've used nanotechnology[br]to develop a real-time translation tool

0:01:08.714,0:01:11.033
that can spy on bacterial communities

0:01:11.057,0:01:13.973
and give us recordings[br]of what bacteria are up to.

0:01:16.123,0:01:17.719
Bacteria live everywhere.

0:01:17.743,0:01:20.072
They're in the soil, on our furniture

0:01:20.096,0:01:21.407
and inside our bodies.

0:01:22.083,0:01:26.622
In fact, 90 percent of all the live cells[br]in this theater are bacterial.

0:01:27.915,0:01:29.514
Some bacteria are good for us;

0:01:29.538,0:01:32.750
they help us digest food[br]or produce antibiotics.

0:01:32.774,0:01:34.866
And some bacteria are bad for us;

0:01:34.890,0:01:36.784
they cause diseases and death.

0:01:37.794,0:01:40.210
To coordinate all[br]the functions bacteria have,

0:01:40.234,0:01:42.306
they have to be able to organize,

0:01:42.330,0:01:44.371
and they do that just like us humans --

0:01:44.395,0:01:45.554
by communicating.

0:01:46.751,0:01:48.226
But instead of using words,

0:01:48.250,0:01:51.192
they use signaling molecules[br]to communicate with each other.

0:01:52.083,0:01:53.340
When bacteria are few,

0:01:53.364,0:01:56.107
the signaling molecules just flow away,

0:01:56.131,0:01:58.636
like the screams of a man[br]alone in the desert.

0:01:59.518,0:02:03.510
But when there are many bacteria,[br]the signaling molecules accumulate,

0:02:03.534,0:02:06.526
and the bacteria start sensing[br]that they're not alone.

0:02:07.309,0:02:08.643
They listen to each other.

0:02:09.459,0:02:12.275
In this way, they keep track[br]of how many they are

0:02:12.299,0:02:15.620
and when they're many enough[br]to initiate a new action.

0:02:16.575,0:02:20.432
And when the signaling molecules[br]have reached a certain threshold,

0:02:20.456,0:02:23.577
all the bacteria sense at once[br]that they need to act

0:02:23.601,0:02:24.919
with the same action.

0:02:25.967,0:02:30.293
So bacterial conversation consists[br]of an initiative and a reaction,

0:02:30.317,0:02:33.389
a production of a molecule[br]and the response to it.

0:02:35.094,0:02:38.434
In my research, I focused on spying[br]on bacterial communities

0:02:38.458,0:02:39.861
inside the human body.

0:02:40.343,0:02:41.588
How does it work?

0:02:42.385,0:02:44.300
We have a sample from a patient.

0:02:44.324,0:02:46.862
It could be a blood or spit sample.

0:02:47.304,0:02:49.841
We shoot electrons into the sample,

0:02:49.865,0:02:53.785
the electrons will interact with any[br]communication molecules present,

0:02:53.809,0:02:56.190
and this interaction[br]will give us information

0:02:56.214,0:02:58.105
on the identity of the bacteria,

0:02:58.129,0:02:59.800
the type of communication

0:02:59.824,0:03:02.117
and how much the bacteria are talking.

0:03:04.269,0:03:06.590
But what is it like[br]when bacteria communicate?

0:03:07.747,0:03:11.507
Before I developed the translation tool,

0:03:11.531,0:03:15.377
my first assumption was that bacteria[br]would have a primitive language,

0:03:15.401,0:03:18.579
like infants that haven't developed[br]words and sentences yet.

0:03:19.208,0:03:22.129
When they laugh, they're happy;[br]when they cry, they're sad.

0:03:22.153,0:03:23.303
Simple as that.

0:03:24.008,0:03:28.123
But bacteria turned out to be nowhere[br]as primitive as I thought they would be.

0:03:28.615,0:03:30.855
A molecule is not just a molecule.

0:03:30.879,0:03:33.633
It can mean different things[br]depending on the context,

0:03:34.404,0:03:37.346
just like the crying of babies[br]can mean different things:

0:03:37.370,0:03:39.140
sometimes the baby is hungry,

0:03:39.164,0:03:40.358
sometimes it's wet,

0:03:40.382,0:03:42.401
sometimes it's hurt or afraid.

0:03:42.425,0:03:44.775
Parents know how to decode those cries.

0:03:45.624,0:03:47.506
And to be a real translation tool,

0:03:47.530,0:03:50.503
it had to be able to decode[br]the signaling molecules

0:03:50.527,0:03:54.588
and translate them[br]depending on the context.

0:03:55.497,0:03:56.648
And who knows?

0:03:56.672,0:03:58.833
Maybe Google Translate[br]will adopt this soon.

0:03:58.857,0:04:01.226
(Laughter)

0:04:02.386,0:04:04.104
Let me give you an example.

0:04:04.128,0:04:07.717
I've brought some bacterial data[br]that can be a bit tricky to understand

0:04:07.741,0:04:08.892
if you're not trained,

0:04:08.916,0:04:10.261
but try to take a look.

0:04:11.548,0:04:13.467
(Laughter)

0:04:14.959,0:04:18.436
Here's a happy bacterial family[br]that has infected a patient.

0:04:20.261,0:04:22.294
Let's call them the Montague family.

0:04:23.920,0:04:27.381
They share resources,[br]they reproduce, and they grow.

0:04:28.294,0:04:30.273
One day, they get a new neighbor,

0:04:32.746,0:04:34.513
bacterial family Capulet.

0:04:34.537,0:04:35.687
(Laughter)

0:04:36.157,0:04:38.947
Everything is fine,[br]as long as they're working together.

0:04:40.377,0:04:43.383
But then something unplanned happens.

0:04:44.449,0:04:48.667
Romeo from Montague has a relationship[br]with Juliet from Capulet.

0:04:48.691,0:04:49.841
(Laughter)

0:04:50.978,0:04:53.873
And yes, they share genetic material.

0:04:53.897,0:04:56.006
(Laughter)

0:04:58.630,0:05:01.381
Now, this gene transfer[br]can be dangerous to the Montagues

0:05:01.405,0:05:05.471
that have the ambition to be the only[br]family in the patient they have infected,

0:05:05.495,0:05:06.919
and sharing genes contributes

0:05:06.943,0:05:09.777
to the Capulets developing[br]resistance to antibiotics.

0:05:11.747,0:05:16.392
So the Montagues start talking internally[br]to get rid of this other family

0:05:16.416,0:05:18.138
by releasing this molecule.

0:05:18.688,0:05:19.838
(Laughter)

0:05:20.700,0:05:22.062
And with subtitles:

0:05:22.372,0:05:23.978
[Let us coordinate an attack.]

0:05:24.002,0:05:25.293
(Laughter)

0:05:25.639,0:05:27.430
Let's coordinate an attack.

0:05:29.148,0:05:32.248
And then everybody at once responds

0:05:32.272,0:05:36.595
by releasing a poison[br]that will kill the other family.

0:05:36.619,0:05:38.387
[Eliminate!]

0:05:40.129,0:05:42.261
(Laughter)

0:05:43.338,0:05:47.731
The Capulets respond[br]by calling for a counterattack.

0:05:47.755,0:05:48.911
[Counterattack!]

0:05:48.935,0:05:50.360
And they have a battle.

0:05:52.090,0:05:56.708
This is a video of real bacteria[br]dueling with swordlike organelles,

0:05:56.732,0:05:58.345
where they try to kill each other

0:05:58.369,0:06:01.207
by literally stabbing[br]and rupturing each other.

0:06:02.784,0:06:06.745
Whoever's family wins this battle[br]becomes the dominant bacteria.

0:06:08.360,0:06:11.639
So what I can do is to detect[br]bacterial conversations

0:06:11.663,0:06:13.695
that lead to different[br]collective behaviors

0:06:13.719,0:06:15.154
like the fight you just saw.

0:06:15.633,0:06:18.550
And what I did was to spy[br]on bacterial communities

0:06:18.574,0:06:20.617
inside the human body

0:06:20.641,0:06:22.357
in patients at a hospital.

0:06:22.737,0:06:25.207
I followed 62 patients in an experiment,

0:06:25.231,0:06:28.979
where I tested the patient samples[br]for one particular infection,

0:06:29.003,0:06:32.332
without knowing the results[br]of the traditional diagnostic test.

0:06:32.356,0:06:36.576
Now, in bacterial diagnostics,

0:06:36.600,0:06:38.581
a sample is smeared out on a plate,

0:06:38.605,0:06:41.729
and if the bacteria grow within five days,

0:06:41.753,0:06:44.117
the patient is diagnosed as infected.

0:06:45.842,0:06:48.661
When I finished the study[br]and I compared the tool results

0:06:48.685,0:06:51.923
to the traditional diagnostic test[br]and the validation test,

0:06:51.947,0:06:53.347
I was shocked.

0:06:53.371,0:06:57.082
It was far more astonishing[br]than I had ever anticipated.

0:06:58.011,0:07:00.150
But before I tell you[br]what the tool revealed,

0:07:00.174,0:07:03.168
I would like to tell you about[br]a specific patient I followed,

0:07:03.192,0:07:04.359
a young girl.

0:07:04.803,0:07:06.253
She had cystic fibrosis,

0:07:06.277,0:07:10.017
a genetic disease that made her lungs[br]susceptible to bacterial infections.

0:07:10.837,0:07:13.233
This girl wasn't a part[br]of the clinical trial.

0:07:13.257,0:07:16.084
I followed her because I knew[br]from her medical record

0:07:16.108,0:07:18.208
that she had never had[br]an infection before.

0:07:19.453,0:07:21.558
Once a month, this girl[br]went to the hospital

0:07:21.582,0:07:24.236
to cough up a sputum sample[br]that she spit in a cup.

0:07:24.916,0:07:28.042
This sample was transferred[br]for bacterial analysis

0:07:28.066,0:07:29.996
at the central laboratory

0:07:30.020,0:07:33.486
so the doctors could act quickly[br]if they discovered an infection.

0:07:34.099,0:07:36.973
And it allowed me to test my device[br]on her samples as well.

0:07:37.355,0:07:40.767
The first two months I measured[br]on her samples, there was nothing.

0:07:41.794,0:07:42.961
But the third month,

0:07:42.985,0:07:45.641
I discovered some bacterial[br]chatter in her sample.

0:07:46.473,0:07:49.585
The bacteria were coordinating[br]to damage her lung tissue.

0:07:50.534,0:07:54.545
But the traditional diagnostics[br]showed no bacteria at all.

0:07:55.711,0:07:57.630
I measured again the next month,

0:07:57.654,0:08:01.282
and I could see that the bacterial[br]conversations became even more aggressive.

0:08:02.167,0:08:04.919
Still, the traditional[br]diagnostics showed nothing.

0:08:06.456,0:08:10.100
My study ended, but a half a year later,[br]I followed up on her status

0:08:10.124,0:08:13.365
to see if the bacteria[br]only I knew about had disappeared

0:08:13.389,0:08:15.404
without medical intervention.

0:08:16.350,0:08:17.500
They hadn't.

0:08:18.020,0:08:20.853
But the girl was now diagnosed[br]with a severe infection

0:08:20.877,0:08:22.195
of deadly bacteria.

0:08:23.511,0:08:27.589
It was the very same bacteria[br]my tool discovered earlier.

0:08:28.537,0:08:31.033
And despite aggressive[br]antibiotic treatment,

0:08:31.057,0:08:33.586
it was impossible[br]to eradicate the infection.

0:08:34.816,0:08:37.898
Doctors deemed that she would not[br]survive her 20s.

0:08:40.404,0:08:42.509
When I measured on this girl's samples,

0:08:42.533,0:08:44.732
my tool was still in the initial stage.

0:08:44.756,0:08:47.455
I didn't even know[br]if my method worked at all,

0:08:47.479,0:08:49.626
therefore I had an agreement[br]with the doctors

0:08:49.650,0:08:51.511
not to tell them what my tool revealed

0:08:51.535,0:08:53.675
in order not to compromise[br]their treatment.

0:08:54.111,0:08:56.914
So when I saw these results[br]that weren't even validated,

0:08:56.938,0:08:58.290
I didn't dare to tell

0:08:58.314,0:09:01.121
because treating a patient[br]without an actual infection

0:09:01.145,0:09:03.755
also has negative[br]consequences for the patient.

0:09:05.092,0:09:06.714
But now we know better,

0:09:06.738,0:09:10.137
and there are many young boys[br]and girls that still can be saved

0:09:11.172,0:09:14.596
because, unfortunately,[br]this scenario happens very often.

0:09:14.620,0:09:16.163
Patients get infected,

0:09:16.187,0:09:19.664
the bacteria somehow don't show[br]on the traditional diagnostic test,

0:09:19.688,0:09:23.540
and suddenly, the infection breaks out[br]in the patient with severe symptoms.

0:09:23.564,0:09:25.722
And at that point, it's already too late.

0:09:27.219,0:09:30.773
The surprising result[br]of the 62 patients I followed

0:09:30.797,0:09:33.340
was that my device[br]caught bacterial conversations

0:09:33.364,0:09:35.580
in more than half of the patient samples

0:09:35.604,0:09:38.535
that were diagnosed as negative[br]by traditional methods.

0:09:39.501,0:09:43.055
In other words, more than half[br]of these patients went home thinking

0:09:43.079,0:09:44.764
they were free from infection,

0:09:44.788,0:09:47.536
although they actually carried[br]dangerous bacteria.

0:09:49.257,0:09:51.554
Inside these wrongly diagnosed patients,

0:09:51.578,0:09:54.598
bacteria were coordinating[br]a synchronized attack.

0:09:55.530,0:09:57.218
They were whispering to each other.

0:09:57.892,0:09:59.523
What I call "whispering bacteria"

0:09:59.547,0:10:02.551
are bacteria that traditional[br]methods cannot diagnose.

0:10:03.383,0:10:07.326
So far, it's only the translation tool[br]that can catch those whispers.

0:10:08.364,0:10:11.767
I believe that the time frame[br]in which bacteria are still whispering

0:10:11.791,0:10:14.778
is a window of opportunity[br]for targeted treatment.

0:10:15.608,0:10:18.741
If the girl had been treated[br]during this window of opportunity,

0:10:18.765,0:10:21.266
it might have been possible[br]to kill the bacteria

0:10:21.290,0:10:22.755
in their initial stage,

0:10:22.779,0:10:24.885
before the infection got out of hand.

0:10:27.131,0:10:31.102
What I experienced with this young girl[br]made me decide to do everything I can

0:10:31.126,0:10:33.331
to push this technology into the hospital.

0:10:34.212,0:10:35.363
Together with doctors,

0:10:35.387,0:10:38.356
I'm already working[br]on implementing this tool in clinics

0:10:38.380,0:10:40.202
to diagnose early infections.

0:10:41.319,0:10:44.563
Although it's still not known[br]how doctors should treat patients

0:10:44.587,0:10:46.400
during the whispering phase,

0:10:46.424,0:10:50.127
this tool can help doctors[br]keep a closer eye on patients in risk.

0:10:50.548,0:10:53.830
It could help them confirm[br]if a treatment had worked or not,

0:10:53.854,0:10:56.618
and it could help answer simple questions:

0:10:56.642,0:10:58.372
Is the patient infected?

0:10:58.396,0:11:00.206
And what are the bacteria up to?

0:11:00.958,0:11:02.745
Bacteria talk,

0:11:02.769,0:11:04.795
they make secret plans,

0:11:04.819,0:11:07.642
and they send confidential[br]information to each other.

0:11:08.253,0:11:10.932
But not only can we catch them whispering,

0:11:10.956,0:11:13.423
we can all learn their secret language

0:11:13.447,0:11:16.326
and become ourselves bacterial whisperers.

0:11:16.973,0:11:18.687
And, as bacteria would say,

0:11:19.656,0:11:22.754
"3-oxo-C12-aniline."

0:11:23.763,0:11:24.931
(Laughter)

0:11:24.955,0:11:26.040
(Applause)

0:11:26.064,0:11:27.248
Thank you.