1
00:00:00,613 --> 00:00:03,001
In 2003,

2
00:00:03,001 --> 00:00:06,032
when we sequenced the human genome,

3
00:00:06,032 --> 00:00:11,126
we thought we would have the answer
to treat many diseases,

4
00:00:11,126 --> 00:00:14,289
but the reality is far from that,

5
00:00:14,289 --> 00:00:16,913
because in addition to our genes,

6
00:00:16,913 --> 00:00:19,619
our environment and lifestyle

7
00:00:19,619 --> 00:00:21,365
could have a significant role

8
00:00:21,365 --> 00:00:23,834
in developing many major diseases.

9
00:00:24,084 --> 00:00:27,680
One example is fatty liver disease,

10
00:00:27,680 --> 00:00:31,291
which is affecting over 20 percent
of the population globally

11
00:00:31,291 --> 00:00:34,504
and it has no treatment
and leads to liver cancer

12
00:00:34,504 --> 00:00:36,085
or liver failure.

13
00:00:36,806 --> 00:00:42,419
So sequencing DNA alone
doesn't give us enough information

14
00:00:42,419 --> 00:00:44,753
to find effective therapeutics.

15
00:00:44,753 --> 00:00:48,559
On the bright side, there are
many other molecules in our body.

16
00:00:48,559 --> 00:00:52,243
In fact, there are
over 100,000 metabolites.

17
00:00:52,243 --> 00:00:57,104
Metabolites are any molecule
that is super-small in their size.

18
00:00:57,104 --> 00:01:02,189
Known examples are glucose,
fructose, fats, cholesterol,

19
00:01:02,189 --> 00:01:04,456
things we hear all the time.

20
00:01:04,456 --> 00:01:07,668
Metabolites are involved
in our metabolism.

21
00:01:07,668 --> 00:01:12,312
They are also downstream of DNA,

22
00:01:12,312 --> 00:01:15,348
so they carry information
from both our genes

23
00:01:15,348 --> 00:01:17,519
as well as lifestyle.

24
00:01:17,519 --> 00:01:23,150
Understanding metabolites is essential
to find treatments for many diseases.

25
00:01:23,150 --> 00:01:26,046
I've always wanted to treat patients.

26
00:01:26,046 --> 00:01:30,032
Despite that, 15 years ago
I left medical school,

27
00:01:30,032 --> 00:01:33,226
as I missed mathematics.

28
00:01:33,226 --> 00:01:36,367
Soon after, I found the coolest thing:

29
00:01:36,367 --> 00:01:39,130
I can use mathematics to study medicine.

30
00:01:41,201 --> 00:01:47,362
Since then, I've been developing
algorithms to analyze biological data.

31
00:01:47,362 --> 00:01:49,710
So it sounded easy:

32
00:01:49,710 --> 00:01:53,254
let's collect data from
all the metabolites in our body,

33
00:01:53,254 --> 00:01:58,355
develop mathematical models to describe
how they are changed in a disease,

34
00:01:58,355 --> 00:02:01,524
and intervene those changes to treat them.

35
00:02:02,719 --> 00:02:06,907
Then I realized why no one
has done this before.

36
00:02:06,907 --> 00:02:10,156
It's extremely difficult.

37
00:02:10,347 --> 00:02:12,783
There are many metabolites in our body.

38
00:02:12,783 --> 00:02:16,062
Each one is different from the other one.

39
00:02:16,062 --> 00:02:19,308
For some metabolites,
we can measure their molecular mass

40
00:02:19,308 --> 00:02:22,032
using mass spectrometry instruments,

41
00:02:22,032 --> 00:02:26,342
but because there could be, like,
10 molecules with the exact same mass,

42
00:02:26,342 --> 00:02:28,143
we don't know exactly what they are,

43
00:02:28,143 --> 00:02:30,986
and if you want to clearly
identify all of them,

44
00:02:30,986 --> 00:02:34,095
you have to do more experiments,
which could take decades

45
00:02:34,095 --> 00:02:35,588
and billions of dollars.

46
00:02:35,588 --> 00:02:38,885
So we developed
an artificial intelligence,

47
00:02:39,290 --> 00:02:42,074
or AI, platform to do that.

48
00:02:42,074 --> 00:02:45,162
We leveraged the growth of biological data

49
00:02:45,162 --> 00:02:49,352
and built a database of any existing
information about metabolites

50
00:02:49,352 --> 00:02:52,525
and their interactions
with other molecules.

51
00:02:52,762 --> 00:02:55,814
We combined all this data
as a mega-network.

52
00:02:55,814 --> 00:02:59,025
then, from tissues or blood of patients,

53
00:02:59,025 --> 00:03:01,905
we measured masses of metabolites

54
00:03:01,905 --> 00:03:05,559
and find the masses
that are changed in a disease.

55
00:03:05,559 --> 00:03:08,756
But, as I mentioned earlier,
we don't know exactly what they are.

56
00:03:08,756 --> 00:03:13,753
Like a molecular mass of 180 could be
the glucose, galactose, or fructose.

57
00:03:13,753 --> 00:03:15,804
They all have the exact same mass,

58
00:03:15,804 --> 00:03:18,009
but different functions in our body.

59
00:03:18,009 --> 00:03:21,540
Our AI algorithm considered
all these ambiguities.

60
00:03:21,540 --> 00:03:24,285
It then mined that mega-network

61
00:03:24,285 --> 00:03:28,724
to find how those metabolic masses
are connected to each other

62
00:03:28,724 --> 00:03:30,639
that result in disease,

63
00:03:30,639 --> 00:03:33,095
and because of the way they are connected,

64
00:03:33,095 --> 00:03:37,054
then we are able to infer
what each metabolite mass is,

65
00:03:37,054 --> 00:03:39,934
like that 180 could be glucose here,

66
00:03:39,934 --> 00:03:44,344
and more importantly to discover
how changes in glucose

67
00:03:44,344 --> 00:03:46,035
and other metabolites

68
00:03:46,035 --> 00:03:47,759
lead to a disease.

69
00:03:47,759 --> 00:03:51,270
This novel understanding
of disease mechanisms

70
00:03:51,270 --> 00:03:55,196
then enable us to discover
effective therapeutics to target that.

71
00:03:55,760 --> 00:03:59,542
So we formed a startup company
to bring this technology to the market

72
00:03:59,542 --> 00:04:01,937
and impacts people's lives.

73
00:04:01,937 --> 00:04:05,474
Now, my team and I at ReviveMed
are working to discover

74
00:04:05,474 --> 00:04:10,657
therapeutics for major diseases
that metabolites are key drivers for,

75
00:04:10,853 --> 00:04:12,476
like fatty liver disease,

76
00:04:12,476 --> 00:04:15,823
because it is caused
by accumulation of fats,

77
00:04:15,823 --> 00:04:17,455
which are types of metabolites,

78
00:04:17,455 --> 00:04:18,400
in the liver,

79
00:04:18,400 --> 00:04:21,967
and as I mentioned earlier
is a huge epidemic with no treatment.

80
00:04:21,967 --> 00:04:24,762
And fatty liver disease
is just one example.

81
00:04:24,762 --> 00:04:28,968
Moving forward, we are going to tackle
hundreds of other diseases

82
00:04:28,968 --> 00:04:30,430
with no treatment,

83
00:04:30,430 --> 00:04:35,009
and by collecting more and more
data about metabolites

84
00:04:35,009 --> 00:04:38,601
and understanding
how changes in metabolites

85
00:04:38,601 --> 00:04:41,127
leads to developing diseases,

86
00:04:41,127 --> 00:04:44,449
our algorithms will get
smarter and smarter

87
00:04:44,449 --> 00:04:48,821
to discover the right therapeutics
for the right patients,

88
00:04:48,821 --> 00:04:52,716
and we will get closer to reach our vision

89
00:04:52,716 --> 00:04:56,383
of saving lives with every line of code.

90
00:04:56,383 --> 00:04:57,754
Thank you.

91
00:04:57,754 --> 00:05:01,581
(Applause)