0:00:00.612,0:00:03.374
For the next 16 minutes,[br]I'm going to take you on a journey

0:00:03.398,0:00:06.484
that is probably[br]the biggest dream of humanity:

0:00:06.508,0:00:08.523
to understand the code of life.

0:00:09.072,0:00:11.815
So for me, everything started[br]many, many years ago

0:00:11.839,0:00:14.562
when I met the first 3D printer.

0:00:14.586,0:00:16.260
The concept was fascinating.

0:00:16.284,0:00:18.306
A 3D printer needs three elements:

0:00:18.330,0:00:22.464
a bit of information, some[br]raw material, some energy,

0:00:22.488,0:00:25.822
and it can produce any object[br]that was not there before.

0:00:26.517,0:00:28.654
I was doing physics,[br]I was coming back home

0:00:28.678,0:00:32.116
and I realized that I actually[br]always knew a 3D printer.

0:00:32.140,0:00:33.476
And everyone does.

0:00:33.500,0:00:34.658
It was my mom.

0:00:34.682,0:00:35.683
(Laughter)

0:00:35.707,0:00:38.121
My mom takes three elements:

0:00:38.145,0:00:42.118
a bit of information, which is between[br]my father and my mom in this case,

0:00:42.142,0:00:46.299
raw elements and energy[br]in the same media, that is food,

0:00:46.323,0:00:48.831
and after several months, produces me.

0:00:48.855,0:00:50.667
And I was not existent before.

0:00:50.691,0:00:54.453
So apart from the shock of my mom[br]discovering that she was a 3D printer,

0:00:54.477,0:00:59.215
I immediately got mesmerized[br]by that piece,

0:00:59.239,0:01:00.956
the first one, the information.

0:01:00.980,0:01:03.231
What amount of information does it take

0:01:03.255,0:01:05.191
to build and assemble a human?

0:01:05.215,0:01:06.789
Is it much? Is it little?

0:01:06.813,0:01:08.993
How many thumb drives can you fill?

0:01:09.017,0:01:11.641
Well, I was studying physics[br]at the beginning

0:01:11.665,0:01:17.262
and I took this approximation of a human[br]as a gigantic Lego piece.

0:01:17.286,0:01:21.071
So, imagine that the building[br]blocks are little atoms

0:01:21.095,0:01:25.748
and there is a hydrogen here,[br]a carbon here, a nitrogen here.

0:01:25.772,0:01:27.343
So in the first approximation,

0:01:27.367,0:01:31.710
if I can list the number of atoms[br]that compose a human being,

0:01:31.734,0:01:33.121
I can build it.

0:01:33.145,0:01:35.174
Now, you can run some numbers

0:01:35.198,0:01:38.475
and that happens to be[br]quite an astonishing number.

0:01:38.499,0:01:41.256
So the number of atoms,

0:01:41.280,0:01:46.035
the file that I will save in my thumb[br]drive to assemble a little baby,

0:01:46.059,0:01:50.726
will actually fill an entire Titanic[br]of thumb drives --

0:01:50.750,0:01:53.468
multiplied 2,000 times.

0:01:53.957,0:01:57.358
This is the miracle of life.

0:01:57.382,0:01:59.994
Every time you see from now on[br]a pregnant lady,

0:02:00.018,0:02:02.874
she's assembling the biggest[br]amount of information

0:02:02.898,0:02:04.454
that you will ever encounter.

0:02:04.478,0:02:07.428
Forget big data, forget[br]anything you heard of.

0:02:07.452,0:02:10.333
This is the biggest amount[br]of information that exists.

0:02:10.357,0:02:14.190
(Applause)

0:02:14.214,0:02:18.858
But nature, fortunately, is much smarter[br]than a young physicist,

0:02:18.882,0:02:22.458
and in four billion years, managed[br]to pack this information

0:02:22.482,0:02:25.187
in a small crystal we call DNA.

0:02:25.605,0:02:29.917
We met it for the first time in 1950[br]when Rosalind Franklin,

0:02:29.941,0:02:31.497
an amazing scientist, a woman,

0:02:31.521,0:02:32.910
took a picture of it.

0:02:32.934,0:02:38.122
But it took us more than 40 years[br]to finally poke inside a human cell,

0:02:38.146,0:02:39.748
take out this crystal,

0:02:39.772,0:02:42.852
unroll it, and read it for the first time.

0:02:43.615,0:02:46.856
The code comes out to be[br]a fairly simple alphabet,

0:02:46.880,0:02:50.652
four letters: A, T, C and G.

0:02:50.676,0:02:54.166
And to build a human,[br]you need three billion of them.

0:02:54.933,0:02:56.112
Three billion.

0:02:56.136,0:02:57.715
How many are three billion?

0:02:57.739,0:03:00.501
It doesn't really make[br]any sense as a number, right?

0:03:00.525,0:03:04.610
So I was thinking how[br]I could explain myself better

0:03:04.634,0:03:07.684
about how big and enormous this code is.

0:03:07.708,0:03:10.762
But there is -- I mean,[br]I'm going to have some help,

0:03:10.786,0:03:14.013
and the best person to help me[br]introduce the code

0:03:14.037,0:03:17.559
is actually the first man[br]to sequence it, Dr. Craig Venter.

0:03:17.583,0:03:20.973
So welcome onstage, Dr. Craig Venter.

0:03:20.997,0:03:27.928
(Applause)

0:03:27.952,0:03:30.208
Not the man in the flesh,

0:03:31.448,0:03:33.793
but for the first time in history,

0:03:33.817,0:03:37.279
this is the genome of a specific human,

0:03:37.303,0:03:41.063
printed page-by-page, letter-by-letter:

0:03:41.087,0:03:45.083
262,000 pages of information,

0:03:45.107,0:03:49.471
450 kilograms, shipped[br]from the United States to Canada

0:03:49.495,0:03:54.338
thanks to Bruno Bowden,[br]Lulu.com, a start-up, did everything.

0:03:54.362,0:03:55.825
It was an amazing feat.

0:03:55.849,0:04:00.146
But this is the visual perception[br]of what is the code of life.

0:04:00.170,0:04:02.648
And now, for the first time,[br]I can do something fun.

0:04:02.672,0:04:05.219
I can actually poke inside it and read.

0:04:05.243,0:04:09.868
So let me take an interesting[br]book ... like this one.

0:04:13.077,0:04:15.611
I have an annotation;[br]it's a fairly big book.

0:04:15.635,0:04:19.362
So just to let you see[br]what is the code of life.

0:04:20.566,0:04:23.957
Thousands and thousands and thousands

0:04:23.981,0:04:26.651
and millions of letters.

0:04:26.675,0:04:29.071
And they apparently make sense.

0:04:29.095,0:04:30.852
Let's get to a specific part.

0:04:31.571,0:04:32.933
Let me read it to you:

0:04:32.957,0:04:33.978
(Laughter)

0:04:34.002,0:04:38.008
"AAG, AAT, ATA."

0:04:38.965,0:04:41.032
To you it sounds like mute letters,

0:04:41.056,0:04:45.097
but this sequence gives[br]the color of the eyes to Craig.

0:04:45.633,0:04:47.565
I'll show you another part of the book.

0:04:47.589,0:04:49.683
This is actually a little[br]more complicated.

0:04:50.983,0:04:53.630
Chromosome 14, book 132:

0:04:53.654,0:04:55.744
(Laughter)

0:04:55.768,0:04:57.045
As you might expect.

0:04:57.069,0:05:00.535
(Laughter)

0:05:02.857,0:05:07.364
"ATT, CTT, GATT."

0:05:08.329,0:05:10.016
This human is lucky,

0:05:10.040,0:05:14.557
because if you miss just[br]two letters in this position --

0:05:14.581,0:05:16.458
two letters of our three billion --

0:05:16.482,0:05:18.501
he will be condemned[br]to a terrible disease:

0:05:18.525,0:05:19.965
cystic fibrosis.

0:05:19.989,0:05:23.402
We have no cure for it,[br]we don't know how to solve it,

0:05:23.426,0:05:27.181
and it's just two letters[br]of difference from what we are.

0:05:27.585,0:05:30.290
A wonderful book, a mighty book,

0:05:31.115,0:05:33.113
a mighty book that helped me understand

0:05:33.137,0:05:35.890
and show you something quite remarkable.

0:05:36.480,0:05:40.915
Every one of you -- what makes[br]me, me and you, you --

0:05:40.939,0:05:43.893
is just about five million of these,

0:05:43.917,0:05:45.145
half a book.

0:05:46.015,0:05:47.678
For the rest,

0:05:47.702,0:05:50.264
we are all absolutely identical.

0:05:51.008,0:05:55.026
Five hundred pages[br]is the miracle of life that you are.

0:05:55.050,0:05:57.581
The rest, we all share it.

0:05:57.605,0:06:00.514
So think about that again[br]when we think that we are different.

0:06:00.538,0:06:02.759
This is the amount that we share.

0:06:03.441,0:06:06.870
So now that I have your attention,

0:06:06.894,0:06:08.253
the next question is:

0:06:08.277,0:06:09.428
How do I read it?

0:06:09.452,0:06:10.961
How do I make sense out of it?

0:06:11.409,0:06:15.649
Well, for however good you can be[br]at assembling Swedish furniture,

0:06:15.673,0:06:19.236
this instruction manual[br]is nothing you can crack in your life.

0:06:19.260,0:06:20.863
(Laughter)

0:06:20.887,0:06:23.999
And so, in 2014, two famous TEDsters,

0:06:24.023,0:06:26.563
Peter Diamandis and Craig Venter himself,

0:06:26.587,0:06:28.514
decided to assemble a new company.

0:06:28.538,0:06:29.950
Human Longevity was born,

0:06:29.974,0:06:31.344
with one mission:

0:06:31.368,0:06:33.229
trying everything we can try

0:06:33.253,0:06:36.012
and learning everything[br]we can learn from these books,

0:06:36.036,0:06:37.741
with one target --

0:06:38.862,0:06:41.663
making real the dream[br]of personalized medicine,

0:06:41.687,0:06:45.454
understanding what things[br]should be done to have better health

0:06:45.478,0:06:47.761
and what are the secrets in these books.

0:06:48.329,0:06:52.579
An amazing team, 40 data scientists[br]and many, many more people,

0:06:52.603,0:06:53.953
a pleasure to work with.

0:06:53.977,0:06:56.230
The concept is actually very simple.

0:06:56.254,0:06:59.412
We're going to use a technology[br]called machine learning.

0:06:59.436,0:07:03.975
On one side, we have genomes --[br]thousands of them.

0:07:03.999,0:07:07.996
On the other side, we collected[br]the biggest database of human beings:

0:07:08.020,0:07:12.316
phenotypes, 3D scan, NMR --[br]everything you can think of.

0:07:12.340,0:07:15.239
Inside there, on these two opposite sides,

0:07:15.263,0:07:17.705
there is the secret of translation.

0:07:17.729,0:07:20.201
And in the middle, we build a machine.

0:07:20.801,0:07:23.186
We build a machine[br]and we train a machine --

0:07:23.210,0:07:26.420
well, not exactly one machine,[br]many, many machines --

0:07:26.444,0:07:30.988
to try to understand and translate[br]the genome in a phenotype.

0:07:31.362,0:07:34.702
What are those letters,[br]and what do they do?

0:07:34.726,0:07:37.473
It's an approach that can[br]be used for everything,

0:07:37.497,0:07:40.490
but using it in genomics[br]is particularly complicated.

0:07:40.514,0:07:43.790
Little by little we grew and we wanted[br]to build different challenges.

0:07:43.814,0:07:46.546
We started from the beginning,[br]from common traits.

0:07:46.570,0:07:49.173
Common traits are comfortable[br]because they are common,

0:07:49.197,0:07:50.381
everyone has them.

0:07:50.405,0:07:52.899
So we started to ask our questions:

0:07:52.923,0:07:54.303
Can we predict height?

0:07:54.985,0:07:57.162
Can we read the books[br]and predict your height?

0:07:57.186,0:07:58.337
Well, we actually can,

0:07:58.361,0:08:00.154
with five centimeters of precision.

0:08:00.178,0:08:03.313
BMI is fairly connected to your lifestyle,

0:08:03.337,0:08:07.201
but we still can, we get in the ballpark,[br]eight kilograms of precision.

0:08:07.225,0:08:08.456
Can we predict eye color?

0:08:08.480,0:08:09.638
Yeah, we can.

0:08:09.662,0:08:10.986
Eighty percent accuracy.

0:08:11.466,0:08:13.324
Can we predict skin color?

0:08:13.348,0:08:15.789
Yeah we can, 80 percent accuracy.

0:08:15.813,0:08:17.153
Can we predict age?

0:08:18.121,0:08:21.860
We can, because apparently,[br]the code changes during your life.

0:08:21.884,0:08:25.166
It gets shorter, you lose pieces,[br]it gets insertions.

0:08:25.190,0:08:27.745
We read the signals, and we make a model.

0:08:28.438,0:08:29.913
Now, an interesting challenge:

0:08:29.937,0:08:31.666
Can we predict a human face?

0:08:33.014,0:08:34.292
It's a little complicated,

0:08:34.316,0:08:37.507
because a human face is scattered[br]among millions of these letters.

0:08:37.531,0:08:40.160
And a human face is not[br]a very well-defined object.

0:08:40.184,0:08:42.235
So, we had to build an entire tier of it

0:08:42.259,0:08:44.969
to learn and teach[br]a machine what a face is,

0:08:44.993,0:08:47.030
and embed and compress it.

0:08:47.054,0:08:49.302
And if you're comfortable[br]with machine learning,

0:08:49.326,0:08:51.610
you understand what the challenge is here.

0:08:52.108,0:08:58.099
Now, after 15 years -- 15 years after[br]we read the first sequence --

0:08:58.123,0:09:01.025
this October, we started[br]to see some signals.

0:09:01.049,0:09:03.504
And it was a very emotional moment.

0:09:03.528,0:09:07.273
What you see here is a subject[br]coming in our lab.

0:09:07.619,0:09:09.547
This is a face for us.

0:09:09.571,0:09:13.202
So we take the real face of a subject,[br]we reduce the complexity,

0:09:13.226,0:09:15.196
because not everything is in your face --

0:09:15.220,0:09:19.006
lots of features and defects[br]and asymmetries come from your life.

0:09:19.030,0:09:22.499
We symmetrize the face,[br]and we run our algorithm.

0:09:23.245,0:09:25.143
The results that I show you right now,

0:09:25.167,0:09:28.539
this is the prediction we have[br]from the blood.

0:09:29.596,0:09:31.120
(Applause)

0:09:31.144,0:09:32.579
Wait a second.

0:09:32.603,0:09:37.295
In these seconds, your eyes are watching,[br]left and right, left and right,

0:09:37.319,0:09:41.249
and your brain wants[br]those pictures to be identical.

0:09:41.273,0:09:43.719
So I ask you to do[br]another exercise, to be honest.

0:09:43.743,0:09:46.030
Please search for the differences,

0:09:46.054,0:09:47.415
which are many.

0:09:47.439,0:09:50.042
The biggest amount of signal[br]comes from gender,

0:09:50.066,0:09:55.267
then there is age, BMI,[br]the ethnicity component of a human.

0:09:55.291,0:09:59.002
And scaling up over that signal[br]is much more complicated.

0:09:59.026,0:10:02.276
But what you see here,[br]even in the differences,

0:10:02.300,0:10:05.895
lets you understand[br]that we are in the right ballpark,

0:10:05.919,0:10:07.267
that we are getting closer.

0:10:07.291,0:10:09.640
And it's already giving you some emotions.

0:10:09.664,0:10:12.367
This is another subject[br]that comes in place,

0:10:12.391,0:10:13.800
and this is a prediction.

0:10:13.824,0:10:18.420
A little smaller face, we didn't get[br]the complete cranial structure,

0:10:18.444,0:10:21.095
but still, it's in the ballpark.

0:10:21.634,0:10:23.858
This is a subject that comes in our lab,

0:10:23.882,0:10:25.325
and this is the prediction.

0:10:26.056,0:10:30.732
So these people have never been seen[br]in the training of the machine.

0:10:30.756,0:10:33.593
These are the so-called "held-out" set.

0:10:33.617,0:10:37.357
But these are people that you will[br]probably never believe.

0:10:37.381,0:10:40.057
We're publishing everything[br]in a scientific publication,

0:10:40.081,0:10:41.232
you can read it.

0:10:41.256,0:10:43.600
But since we are onstage,[br]Chris challenged me.

0:10:43.624,0:10:47.250
I probably exposed myself[br]and tried to predict

0:10:47.274,0:10:50.105
someone that you might recognize.

0:10:50.470,0:10:54.895
So, in this vial of blood --[br]and believe me, you have no idea

0:10:54.919,0:10:57.799
what we had to do to have[br]this blood now, here --

0:10:57.823,0:11:01.724
in this vial of blood is the amount[br]of biological information

0:11:01.748,0:11:04.025
that we need to do a full genome sequence.

0:11:04.049,0:11:06.119
We just need this amount.

0:11:06.528,0:11:09.733
We ran this sequence,[br]and I'm going to do it with you.

0:11:09.757,0:11:13.736
And we start to layer up[br]all the understanding we have.

0:11:13.760,0:11:17.110
In the vial of blood,[br]we predicted he's a male.

0:11:17.134,0:11:18.498
And the subject is a male.

0:11:18.996,0:11:21.434
We predict that he's a meter and 76 cm.

0:11:21.458,0:11:23.850
The subject is a meter and 77 cm.

0:11:23.874,0:11:27.984
So, we predicted that he's 76;[br]the subject is 82.

0:11:28.701,0:11:31.333
We predict his age, 38.

0:11:31.357,0:11:33.261
The subject is 35.

0:11:33.851,0:11:35.975
We predict his eye color.

0:11:36.824,0:11:38.035
Too dark.

0:11:38.059,0:11:39.614
We predict his skin color.

0:11:40.026,0:11:41.436
We are almost there.

0:11:41.899,0:11:43.272
That's his face.

0:11:45.172,0:11:48.441
Now, the reveal moment:

0:11:48.465,0:11:50.235
the subject is this person.

0:11:50.259,0:11:52.194
(Laughter)

0:11:52.218,0:11:54.276
And I did it intentionally.

0:11:54.300,0:11:57.992
I am a very particular[br]and peculiar ethnicity.

0:11:58.016,0:12:00.966
Southern European, Italians --[br]they never fit in models.

0:12:00.990,0:12:06.120
And it's particular -- that ethnicity[br]is a complex corner case for our model.

0:12:06.144,0:12:07.653
But there is another point.

0:12:07.677,0:12:11.154
So, one of the things that we use[br]a lot to recognize people

0:12:11.178,0:12:12.900
will never be written in the genome.

0:12:12.924,0:12:15.241
It's our free will, it's how I look.

0:12:15.265,0:12:18.494
Not my haircut in this case,[br]but my beard cut.

0:12:18.518,0:12:22.071
So I'm going to show you, I'm going to,[br]in this case, transfer it --

0:12:22.095,0:12:24.860
and this is nothing more[br]than Photoshop, no modeling --

0:12:24.884,0:12:26.597
the beard on the subject.

0:12:26.621,0:12:30.093
And immediately, we get[br]much, much better in the feeling.

0:12:30.955,0:12:33.664
So, why do we do this?

0:12:35.938,0:12:41.078
We certainly don't do it[br]for predicting height

0:12:41.102,0:12:43.474
or taking a beautiful picture[br]out of your blood.

0:12:44.390,0:12:48.408
We do it because the same technology[br]and the same approach,

0:12:48.432,0:12:50.952
the machine learning of this code,

0:12:50.976,0:12:54.113
is helping us to understand how we work,

0:12:54.137,0:12:55.623
how your body works,

0:12:55.647,0:12:57.312
how your body ages,

0:12:57.336,0:13:00.105
how disease generates in your body,

0:13:00.129,0:13:03.101
how your cancer grows and develops,

0:13:03.125,0:13:04.908
how drugs work

0:13:04.932,0:13:07.246
and if they work on your body.

0:13:07.713,0:13:09.380
This is a huge challenge.

0:13:09.894,0:13:11.532
This is a challenge that we share

0:13:11.556,0:13:14.135
with thousands of other[br]researchers around the world.

0:13:14.159,0:13:16.381
It's called personalized medicine.

0:13:17.125,0:13:20.585
It's the ability to move[br]from a statistical approach

0:13:20.609,0:13:22.641
where you're a dot in the ocean,

0:13:22.665,0:13:24.478
to a personalized approach,

0:13:24.502,0:13:26.687
where we read all these books

0:13:26.711,0:13:29.575
and we get an understanding[br]of exactly how you are.

0:13:30.260,0:13:33.622
But it is a particularly[br]complicated challenge,

0:13:33.646,0:13:37.644
because of all these books, as of today,

0:13:37.668,0:13:40.310
we just know probably two percent:

0:13:41.027,0:13:44.680
four books of more than 175.

0:13:46.021,0:13:49.227
And this is not the topic of my talk,

0:13:50.145,0:13:52.743
because we will learn more.

0:13:53.378,0:13:56.047
There are the best minds[br]in the world on this topic.

0:13:57.048,0:13:58.882
The prediction will get better,

0:13:58.906,0:14:01.159
the model will get more precise.

0:14:01.183,0:14:03.041
And the more we learn,

0:14:03.065,0:14:07.895
the more we will[br]be confronted with decisions

0:14:07.919,0:14:10.940
that we never had to face before

0:14:10.964,0:14:12.399
about life,

0:14:12.423,0:14:14.097
about death,

0:14:14.121,0:14:15.724
about parenting.

0:14:20.626,0:14:25.372
So, we are touching the very[br]inner detail on how life works.

0:14:26.118,0:14:29.276
And it's a revolution[br]that cannot be confined

0:14:29.300,0:14:31.959
in the domain of science or technology.

0:14:32.960,0:14:35.204
This must be a global conversation.

0:14:35.798,0:14:41.015
We must start to think of the future[br]we're building as a humanity.

0:14:41.039,0:14:45.103
We need to interact with creatives,[br]with artists, with philosophers,

0:14:45.127,0:14:46.637
with politicians.

0:14:46.661,0:14:47.819
Everyone is involved,

0:14:47.843,0:14:50.668
because it's the future of our species.

0:14:51.273,0:14:55.241
Without fear, but with the understanding

0:14:55.265,0:14:59.136
that the decisions[br]that we make in the next year

0:14:59.160,0:15:02.949
will change the course of history forever.

0:15:03.732,0:15:04.892
Thank you.

0:15:04.916,0:15:15.075
(Applause)