1
00:00:01,507 --> 00:00:03,396
In 2003,

2
00:00:03,420 --> 00:00:06,333
when we sequenced the human genome,

3
00:00:06,357 --> 00:00:10,279
we thought we would have the answer
to treat many diseases.

4
00:00:10,974 --> 00:00:13,631
But the reality is far from that,

5
00:00:14,782 --> 00:00:16,703
because in addition to our genes,

6
00:00:16,727 --> 00:00:21,297
our environment and lifestyle
could have a significant role

7
00:00:21,321 --> 00:00:23,869
in developing many major diseases.

8
00:00:23,893 --> 00:00:27,473
One example is fatty liver disease,

9
00:00:27,497 --> 00:00:31,580
which is affecting over 20 percent
of the population globally,

10
00:00:31,604 --> 00:00:34,638
and it has no treatment
and leads to liver cancer

11
00:00:34,662 --> 00:00:36,085
or liver failure.

12
00:00:37,517 --> 00:00:42,261
So sequencing DNA alone
doesn't give us enough information

13
00:00:42,285 --> 00:00:44,517
to find effective therapeutics.

14
00:00:44,541 --> 00:00:48,297
On the bright side, there are
many other molecules in our body.

15
00:00:48,321 --> 00:00:52,301
In fact, there are
over 100,000 metabolites.

16
00:00:52,325 --> 00:00:56,621
Metabolites are any molecule
that is supersmall in their size.

17
00:00:57,193 --> 00:01:02,165
Known examples are glucose,
fructose, fats, cholesterol --

18
00:01:02,189 --> 00:01:03,699
things we hear all the time.

19
00:01:04,273 --> 00:01:07,256
Metabolites are involved
in our metabolism.

20
00:01:08,066 --> 00:01:12,094
They are also downstream of DNA,

21
00:01:12,118 --> 00:01:17,200
so they carry information
from both our genes as well as lifestyle.

22
00:01:17,224 --> 00:01:22,873
Understanding metabolites is essential
to find treatments for many diseases.

23
00:01:22,897 --> 00:01:25,109
I've always wanted to treat patients.

24
00:01:25,934 --> 00:01:29,792
Despite that, 15 years ago,
I left medical school,

25
00:01:29,816 --> 00:01:31,781
as I missed mathematics.

26
00:01:33,019 --> 00:01:35,955
Soon after, I found the coolest thing:

27
00:01:36,692 --> 00:01:39,455
I can use mathematics to study medicine.

28
00:01:41,026 --> 00:01:46,239
Since then, I've been developing
algorithms to analyze biological data.

29
00:01:47,092 --> 00:01:49,375
So, it sounded easy:

30
00:01:49,399 --> 00:01:53,000
let's collect data from all
the metabolites in our body,

31
00:01:53,024 --> 00:01:58,152
develop mathematical models to describe
how they are changed in a disease

32
00:01:58,176 --> 00:02:01,164
and intervene in those
changes to treat them.

33
00:02:02,488 --> 00:02:05,960
Then I realized why no one
has done this before:

34
00:02:07,230 --> 00:02:08,917
it's extremely difficult.

35
00:02:08,941 --> 00:02:10,028
(Laughter)

36
00:02:10,052 --> 00:02:12,464
There are many metabolites in our body.

37
00:02:12,783 --> 00:02:15,283
Each one is different from the other one.

38
00:02:15,307 --> 00:02:19,035
For some metabolites,
we can measure their molecular mass

39
00:02:19,059 --> 00:02:21,652
using mass spectrometry instruments.

40
00:02:21,676 --> 00:02:26,069
But because there could be, like,
10 molecules with the exact same mass,

41
00:02:26,093 --> 00:02:27,900
we don't know exactly what they are,

42
00:02:27,924 --> 00:02:30,698
and if you want to clearly
identify all of them,

43
00:02:30,722 --> 00:02:33,826
you have to do more experiments,
which could take decades

44
00:02:33,850 --> 00:02:35,564
and billions of dollars.

45
00:02:36,207 --> 00:02:41,770
So we developed an artificial
intelligence, or AI, platform, to do that.

46
00:02:41,794 --> 00:02:44,638
We leveraged the growth of biological data

47
00:02:44,662 --> 00:02:49,086
and built a database of any existing
information about metabolites

48
00:02:49,110 --> 00:02:52,238
and their interactions
with other molecules.

49
00:02:52,262 --> 00:02:55,686
We combined all this data
as a meganetwork.

50
00:02:55,710 --> 00:02:59,106
Then, from tissues or blood of patients,

51
00:02:59,130 --> 00:03:01,881
we measure masses of metabolites

52
00:03:01,905 --> 00:03:05,164
and find the masses
that are changed in a disease.

53
00:03:05,188 --> 00:03:08,378
But, as I mentioned earlier,
we don't know exactly what they are.

54
00:03:08,402 --> 00:03:13,537
A molecular mass of 180 could be
either the glucose, galactose or fructose.

55
00:03:13,561 --> 00:03:15,580
They all have the exact same mass

56
00:03:15,604 --> 00:03:17,691
but different functions in our body.

57
00:03:17,715 --> 00:03:21,302
Our AI algorithm considered
all these ambiguities.

58
00:03:21,326 --> 00:03:24,062
It then mined that meganetwork

59
00:03:24,086 --> 00:03:28,439
to find how those metabolic masses
are connected to each other

60
00:03:28,463 --> 00:03:30,421
that result in disease.

61
00:03:30,445 --> 00:03:32,683
And because of the way they are connected,

62
00:03:32,707 --> 00:03:37,030
then we are able to infer
what each metabolite mass is,

63
00:03:37,054 --> 00:03:39,978
like that 180 could be glucose here,

64
00:03:40,002 --> 00:03:42,553
and, more importantly, to discover

65
00:03:42,577 --> 00:03:45,944
how changes in glucose
and other metabolites

66
00:03:45,968 --> 00:03:47,473
lead to a disease.

67
00:03:47,497 --> 00:03:50,492
This novel understanding
of disease mechanisms

68
00:03:50,516 --> 00:03:55,008
then enable us to discover
effective therapeutics to target that.

69
00:03:55,601 --> 00:03:59,446
So we formed a start-up company
to bring this technology to the market

70
00:03:59,470 --> 00:04:01,275
and impact people's lives.

71
00:04:01,722 --> 00:04:05,267
Now my team and I at ReviveMed
are working to discover

72
00:04:05,291 --> 00:04:10,396
therapeutics for major diseases
that metabolites are key drivers for,

73
00:04:10,420 --> 00:04:12,317
like fatty liver disease,

74
00:04:12,341 --> 00:04:15,265
because it is caused
by accumulation of fats,

75
00:04:15,289 --> 00:04:17,762
which are types
of metabolites in the liver.

76
00:04:17,786 --> 00:04:21,726
As I mentioned earlier,
it's a huge epidemic with no treatment.

77
00:04:21,750 --> 00:04:24,474
And fatty liver disease
is just one example.

78
00:04:24,498 --> 00:04:28,676
Moving forward, we are going to tackle
hundreds of other diseases

79
00:04:28,700 --> 00:04:30,193
with no treatment.

80
00:04:30,217 --> 00:04:34,771
And by collecting more and more
data about metabolites

81
00:04:34,795 --> 00:04:38,339
and understanding
how changes in metabolites

82
00:04:38,363 --> 00:04:40,765
leads to developing diseases,

83
00:04:40,789 --> 00:04:44,278
our algorithms will get
smarter and smarter

84
00:04:44,302 --> 00:04:48,498
to discover the right therapeutics
for the right patients.

85
00:04:48,522 --> 00:04:52,292
And we will get closer to reach our vision

86
00:04:52,316 --> 00:04:56,179
of saving lives with every line of code.

87
00:04:56,203 --> 00:04:57,524
Thank you.

88
00:04:57,548 --> 00:05:01,375
(Applause)