TED2019
Leila Pirhaji: The medical potential of AI and metabolites
萊拉·皮爾哈吉: 人工智慧與代謝物的醫學潛力
Filmed:
Readability: 5.5
1,210,569 views
許多疾病由代謝物引起。代謝物是身體中的微小分子物質，像是脂肪、葡萄糖和膽固醇。但是我們無法分辨出它們究竟是什麼，也不知道它們如何起作用。生技創業家和 TED 研究員萊拉·皮爾哈吉恩（Leila Pirhaji），分享她如何建立一套人工智能網絡，用來分析代謝物型態，從中更瞭解疾病發生的機制，並且找到更有效的治療方法。
Leila Pirhaji - Biotech entrepreneur
Leila Pirhaji uses artificial intelligence for drug discovery and the treatment of metabolic diseases. Full bio
Leila Pirhaji uses artificial intelligence for drug discovery and the treatment of metabolic diseases. Full bio
Double-click the English transcript below to play the video.
00:13
In 2003,
0
1507
1889
在 2003 年，
00:15
when we sequenced the human genome,
1
3420
2913
當我們為人類的基因組定序時，
00:18
we thought we would have the answer
to treat many diseases.
to treat many diseases.
2
6357
3922
我們以為會找到許多疾病的治療方法。
00:22
But the reality is far from that,
3
10974
2657
但實際情形卻遠非如此，
00:26
because in addition to our genes,
4
14782
1921
因為除了我們的基因之外，
00:28
our environment and lifestyle
could have a significant role
could have a significant role
5
16727
4570
生活環境和生活作息
也是引發重大疾病的關鍵因素。
00:33
in developing many major diseases.
6
21321
2548
00:35
One example is fatty liver disease,
7
23893
3580
以脂肪肝疾病爲例，
00:39
which is affecting over 20 percent
of the population globally,
of the population globally,
8
27497
4083
全球超過 20% 的人口
受此疾病影響，
受此疾病影響，
00:43
and it has no treatment
and leads to liver cancer
and leads to liver cancer
9
31604
3034
目前沒有任何治療方法
而且最後可發展為肝癌，
而且最後可發展為肝癌，
00:46
or liver failure.
10
34662
1423
或是肝臟衰竭。
00:49
So sequencing DNA alone
doesn't give us enough information
doesn't give us enough information
11
37517
4744
所以只靠基因定序
並不能給我們足夠的訊息，
並不能給我們足夠的訊息，
00:54
to find effective therapeutics.
12
42285
2232
找出有效的治療方法。
00:56
On the bright side, there are
many other molecules in our body.
many other molecules in our body.
13
44541
3756
好消息是，我們身體裡
還有許多其他的分子，
還有許多其他的分子，
01:00
In fact, there are
over 100,000 metabolites.
over 100,000 metabolites.
14
48321
3980
事實上，我們身體
有超過十萬的代謝物。
有超過十萬的代謝物。
01:04
Metabolites are any molecule
that is supersmall in their size.
that is supersmall in their size.
15
52325
4296
代謝物是體積超級小的分子，
01:09
Known examples are glucose,
fructose, fats, cholesterol --
fructose, fats, cholesterol --
16
57193
4972
已知的例子包括，
葡萄糖、果糖、脂肪、膽固醇——
葡萄糖、果糖、脂肪、膽固醇——
01:14
things we hear all the time.
17
62189
1510
我們時常聽到的這些東西。
01:16
Metabolites are involved
in our metabolism.
in our metabolism.
18
64273
2983
代謝物會參與新陳代謝活動，
01:20
They are also downstream of DNA,
19
68066
4028
它們也是 DNA 的後段，
01:24
so they carry information
from both our genes as well as lifestyle.
from both our genes as well as lifestyle.
20
72118
5082
所以它們帶著基因訊息
也透露出我們的生活作息。
也透露出我們的生活作息。
01:29
Understanding metabolites is essential
to find treatments for many diseases.
to find treatments for many diseases.
21
77224
5649
要找出許多疾病的治療方法
就有必要瞭解代謝物，
就有必要瞭解代謝物，
01:34
I've always wanted to treat patients.
22
82897
2212
我一直都想要醫治好病人，
01:37
Despite that, 15 years ago,
I left medical school,
I left medical school,
23
85934
3858
但是十五年前，
因爲傾心於數學而離開了醫學院。
01:41
as I missed mathematics.
24
89816
1965
01:45
Soon after, I found the coolest thing:
25
93019
2936
不久，我發現最酷的事情是：
01:48
I can use mathematics to study medicine.
26
96692
2763
我可以用數學來研究醫學，
01:53
Since then, I've been developing
algorithms to analyze biological data.
algorithms to analyze biological data.
27
101026
5213
從那時起，我就一直開發
演算法用來分析生物數據。
演算法用來分析生物數據。
01:59
So, it sounded easy:
28
107092
2283
這聽起來很簡單：
02:01
let's collect data from all
the metabolites in our body,
the metabolites in our body,
29
109399
3601
我們收集身體中所有代謝物的數據，
02:05
develop mathematical models to describe
how they are changed in a disease
how they are changed in a disease
30
113024
5128
然後開發數學模型來描述
它們在疾病中如何變化，
它們在疾病中如何變化，
02:10
and intervene in those
changes to treat them.
changes to treat them.
31
118176
2988
並且干預這些變化來進行治療。
02:14
Then I realized why no one
has done this before:
has done this before:
32
122488
3472
然後，我明白為什麼之前
沒有人做過這件事了：
沒有人做過這件事了：
02:19
it's extremely difficult.
33
127230
1687
因為這實在太困難了。
02:20
(Laughter)
34
128941
1087
（笑聲）
02:22
There are many metabolites in our body.
35
130052
2412
我們身體中有太多代謝物了，
02:24
Each one is different from the other one.
36
132783
2500
每一個都不盡相同。
02:27
For some metabolites,
we can measure their molecular mass
we can measure their molecular mass
37
135307
3728
針對一些代謝物，
我們能夠用質譜儀
來測量它們的分子量。
來測量它們的分子量。
02:31
using mass spectrometry instruments.
38
139059
2593
02:33
But because there could be, like,
10 molecules with the exact same mass,
10 molecules with the exact same mass,
39
141676
4393
但是具有完全相同的
分子量可能有十種之多，
分子量可能有十種之多，
02:38
we don't know exactly what they are,
40
146093
1807
所以無法知道它們確切是什麼東西，
02:39
and if you want to clearly
identify all of them,
identify all of them,
41
147924
2774
假如要清楚辨識所有代謝物，
02:42
you have to do more experiments,
which could take decades
which could take decades
42
150722
3104
必須要做更多的實驗，
那有可能要花上數十年的時間，
那有可能要花上數十年的時間，
02:45
and billions of dollars.
43
153850
1714
還要耗費幾十億美元。
02:48
So we developed an artificial
intelligence, or AI, platform, to do that.
intelligence, or AI, platform, to do that.
44
156207
5563
因此，我們開發了一種
人工智慧來做這事。
人工智慧來做這事。
02:53
We leveraged the growth of biological data
45
161794
2844
我們利用生物數據的增長，
02:56
and built a database of any existing
information about metabolites
information about metabolites
46
164662
4424
然後建立一個資料庫
裡面有代謝物的相關訊息，
裡面有代謝物的相關訊息，
03:01
and their interactions
with other molecules.
with other molecules.
47
169110
3128
包含代謝物與其他分子
相互作用的訊息，
相互作用的訊息，
03:04
We combined all this data
as a meganetwork.
as a meganetwork.
48
172262
3424
我們把所有數據組合成一個巨大網絡。
03:07
Then, from tissues or blood of patients,
49
175710
3396
接著，從患者的器官組織或是血液中，
03:11
we measure masses of metabolites
50
179130
2751
我們測量到代謝物的分子量，
03:13
and find the masses
that are changed in a disease.
that are changed in a disease.
51
181905
3259
並且尋找因疾病
而產生變化的代謝物質量。
而產生變化的代謝物質量。
03:17
But, as I mentioned earlier,
we don't know exactly what they are.
we don't know exactly what they are.
52
185188
3190
但是，正如我稍早提過，
我們無法確切知道它們是什麼，
我們無法確切知道它們是什麼，
03:20
A molecular mass of 180 could be
either the glucose, galactose or fructose.
either the glucose, galactose or fructose.
53
188402
5135
分子量為 180 可能是葡萄糖，
不然就是半乳糖或是果糖，
不然就是半乳糖或是果糖，
03:25
They all have the exact same mass
54
193561
2019
它們都擁有相同的質量，
03:27
but different functions in our body.
55
195604
2087
但在身體中有著不同的功能。
03:29
Our AI algorithm considered
all these ambiguities.
all these ambiguities.
56
197715
3587
我們的人工智慧演算考慮到
這些含糊不清的情形，
這些含糊不清的情形，
03:33
It then mined that meganetwork
57
201326
2736
它會在巨大網絡中挖掘數據，
03:36
to find how those metabolic masses
are connected to each other
are connected to each other
58
204086
4353
找出那些代謝物如何相互連結，
03:40
that result in disease.
59
208463
1958
才會導致疾病的發生。
03:42
And because of the way they are connected,
60
210445
2238
而且因為它們連接的方式，
03:44
then we are able to infer
what each metabolite mass is,
what each metabolite mass is,
61
212707
4323
我們得以推斷出
每個代謝物的分子量是多少。
每個代謝物的分子量是多少。
03:49
like that 180 could be glucose here,
62
217054
2924
在這裡，分子量 180 的可能是葡萄糖。
03:52
and, more importantly, to discover
63
220002
2551
而且更重要的是，
發現葡萄糖和其他代謝物的變化
03:54
how changes in glucose
and other metabolites
and other metabolites
64
222577
3367
03:57
lead to a disease.
65
225968
1505
如何引發疾病。
03:59
This novel understanding
of disease mechanisms
of disease mechanisms
66
227497
2995
這種針對疾病機制的新穎理解，
04:02
then enable us to discover
effective therapeutics to target that.
effective therapeutics to target that.
67
230516
4492
讓我們能夠針對疾病
找出有效的治療方法。
找出有效的治療方法。
04:07
So we formed a start-up company
to bring this technology to the market
to bring this technology to the market
68
235601
3845
所以我們成立了一家新創公司
將這項技術帶入市場，
將這項技術帶入市場，
04:11
and impact people's lives.
69
239470
1805
對大家的生活帶來正面影響，
04:13
Now my team and I at ReviveMed
are working to discover
are working to discover
70
241722
3545
現在我和團隊
在 ReviveMed 生技公司
在 ReviveMed 生技公司
正利用代謝物
努力尋找重大疾病的療法，
努力尋找重大疾病的療法，
04:17
therapeutics for major diseases
that metabolites are key drivers for,
that metabolites are key drivers for,
71
245291
5105
04:22
like fatty liver disease,
72
250420
1897
像是脂肪肝疾病，
04:24
because it is caused
by accumulation of fats,
by accumulation of fats,
73
252341
2924
這是由於脂肪的堆積引起。
04:27
which are types
of metabolites in the liver.
of metabolites in the liver.
74
255289
2473
而脂肪是肝臟中
不同類型的代謝物組成，
不同類型的代謝物組成，
04:29
As I mentioned earlier,
it's a huge epidemic with no treatment.
it's a huge epidemic with no treatment.
75
257786
3940
我稍早提到這種重大疾病
目前沒有任何治療方式，
目前沒有任何治療方式，
04:33
And fatty liver disease
is just one example.
is just one example.
76
261750
2724
脂肪肝疾病只是其中一個例子，
04:36
Moving forward, we are going to tackle
hundreds of other diseases
hundreds of other diseases
77
264498
4178
我們接著要解決
其他數百種目前尚無治療方式的疾病。
04:40
with no treatment.
78
268700
1493
04:42
And by collecting more and more
data about metabolites
data about metabolites
79
270217
4554
藉著搜集更多的代謝物數據，
04:46
and understanding
how changes in metabolites
how changes in metabolites
80
274795
3544
並且瞭解這些代謝物的變化
04:50
leads to developing diseases,
81
278363
2402
如何引發疾病。
04:52
our algorithms will get
smarter and smarter
smarter and smarter
82
280789
3489
我們的演算法會變得愈來愈聰明，
04:56
to discover the right therapeutics
for the right patients.
for the right patients.
83
284302
4196
幫助病患找出正確的治療方法。
05:00
And we will get closer to reach our vision
84
288522
3770
而且我們能夠利用每條基因碼
05:04
of saving lives with every line of code.
85
292316
3863
一步步達成拯救生命的願景。
05:08
Thank you.
86
296203
1321
謝謝大家。
05:09
(Applause)
87
297548
3827
（掌聲）
ABOUT THE SPEAKER
Leila Pirhaji - Biotech entrepreneurLeila Pirhaji uses artificial intelligence for drug discovery and the treatment of metabolic diseases.
Why you should listen
Leila Pirhaji is the founder of ReviveMed, an AI platform that can quickly and inexpensively characterize large numbers of metabolites (small molecules like glucose or cholesterol) from the blood, urine and tissues of patients. This allows for the detection of molecular mechanisms that lead to disease and the discovery of drugs that target these disease mechanisms.
Leila Pirhaji | Speaker | TED.com