[00:00.000 --> 00:02.260]  Okay, we're live. Take it away.
[00:03.800 --> 00:05.840]  Hi. Hi there.
[00:06.620 --> 00:07.760]  Can they see us?
[00:07.760 --> 00:10.480]  Yes, I guess so. Let's see whether they can...
[00:10.480 --> 00:11.920]  You are live. You're live.
[00:11.920 --> 00:14.460]  We are live. Okay, so here we are.
[00:14.460 --> 00:16.460]  Hi everyone. Hi. I'm Noa.
[00:16.460 --> 00:17.820]  This is Lawrence.
[00:17.860 --> 00:19.600]  Nice to meet you. We are here
[00:20.180 --> 00:22.380]  streaming from the Weissman Institute of Science
[00:22.380 --> 00:23.800]  in Israel. And
[00:23.800 --> 00:25.800]  together we are going to present
[00:26.700 --> 00:27.180]  the
[00:28.500 --> 00:29.460]  comparison between
[00:29.460 --> 00:31.160]  cybersecurity and biology.
[00:31.160 --> 00:32.340]  I hope you are ready.
[00:32.800 --> 00:35.180]  Well, I'm ready, so I guess the stage is yours, Noa.
[00:35.180 --> 00:35.980]  Okay.
[00:37.940 --> 00:39.260]  So, nice to
[00:39.260 --> 00:40.620]  meet you. I will just
[00:42.720 --> 00:43.200]  present...
[00:43.200 --> 00:45.360]  Okay. Sorry.
[00:45.720 --> 00:47.100]  Okay, so our agenda
[00:47.100 --> 00:49.080]  for today... I have two
[00:49.080 --> 00:51.020]  targets with my presentation.
[00:51.020 --> 00:53.180]  First of all, I'll show you how beautiful and complex
[00:53.180 --> 00:55.100]  biology is. And the second
[00:55.100 --> 00:57.300]  part will be to make you jealous.
[00:57.600 --> 00:59.120]  So, who am I?
[00:59.120 --> 01:01.080]  My name is Noa. I'm a master
[01:01.080 --> 01:03.320]  student at the Weissman Institute of Science.
[01:03.320 --> 01:04.720]  I'm a junior scientist in
[01:05.780 --> 01:06.380]  a lab
[01:07.240 --> 01:08.820]  that is doing research
[01:08.820 --> 01:11.820]  specifically in microbiome, but also diabetes.
[01:12.040 --> 01:13.080]  And I'm a former
[01:13.080 --> 01:15.600]  cybersecurity researcher at
[01:15.600 --> 01:16.800]  Checkpoint.
[01:18.480 --> 01:19.380]  So,
[01:19.620 --> 01:21.880]  until a year ago,
[01:21.880 --> 01:23.380]  I was a
[01:23.380 --> 01:25.300]  reverser in Checkpoint. My job
[01:25.300 --> 01:26.380]  was to reverse...
[01:27.500 --> 01:30.320]  I'm sure that you all know about this job,
[01:30.320 --> 01:31.600]  to reverse engineering malware,
[01:31.900 --> 01:33.320]  and then to implement
[01:34.140 --> 01:35.260]  the defensive
[01:35.260 --> 01:37.300]  coding into Checkpoint
[01:38.240 --> 01:39.360]  antivirus.
[01:39.960 --> 01:41.260]  So, you know, I thought
[01:41.260 --> 01:44.100]  I'm a pretty smart girl, after all.
[01:44.160 --> 01:45.260]  I even did
[01:45.260 --> 01:47.340]  some fuzzing stuff. I found some
[01:47.340 --> 01:48.300]  vulnerabilities.
[01:50.260 --> 01:51.340]  You know, I
[01:51.340 --> 01:53.780]  built myself up as a
[01:53.780 --> 01:55.640]  hacker, as a researcher
[01:55.640 --> 01:57.860]  in Checkpoint for many years.
[02:01.040 --> 02:02.440]  And then...
[02:02.440 --> 02:03.700]  But I was always
[02:03.700 --> 02:05.820]  very curious about complex
[02:05.820 --> 02:08.020]  themes, not only malwares, but also
[02:08.020 --> 02:10.300]  about biology.
[02:10.460 --> 02:11.760]  And a year ago, I decided
[02:11.760 --> 02:14.940]  to come here and to become
[02:14.940 --> 02:17.260]  a researcher in biology,
[02:17.260 --> 02:18.660]  specifically about
[02:18.660 --> 02:19.320]  the new area in
[02:19.320 --> 02:20.760]  microbiome.
[02:20.860 --> 02:23.240]  So I thought, you know, I'm so cool,
[02:23.240 --> 02:24.500]  and I'm doing reverse engineering
[02:25.440 --> 02:27.400]  for so many years.
[02:27.460 --> 02:29.120]  So, you know, all of the
[02:29.120 --> 02:31.400]  bacteria and the COVID-19 and whatever
[02:31.940 --> 02:33.120]  comes up,
[02:33.120 --> 02:34.700]  all of them
[02:34.700 --> 02:36.040]  are going to be
[02:37.020 --> 02:38.960]  scared of me, because now
[02:38.960 --> 02:41.040]  there is a good reverser and a good hacker
[02:41.040 --> 02:42.780]  that is coming into the field.
[02:42.880 --> 02:44.720]  But then I found out that
[02:46.000 --> 02:47.440]  apparently it's not as...
[02:47.440 --> 02:49.540]  I think that we all...
[02:49.540 --> 02:51.540]  I think that it will be really
[02:51.540 --> 02:53.520]  easy for me to prove my point that it's
[02:53.520 --> 02:55.440]  not as easy as we can...
[02:56.080 --> 02:57.440]  as we thought,
[02:57.440 --> 03:00.120]  until not very far.
[03:00.700 --> 03:01.840]  You can
[03:01.840 --> 03:03.540]  see, like, what happened when
[03:03.540 --> 03:05.360]  just one bacteria just
[03:05.360 --> 03:07.300]  did some
[03:08.100 --> 03:09.300]  evolution and become
[03:09.300 --> 03:10.660]  just a little bit
[03:11.380 --> 03:13.560]  more dangerous than before.
[03:13.560 --> 03:15.520]  Like, the COVID
[03:15.520 --> 03:17.560]  is doing evolution
[03:17.560 --> 03:19.520]  for many years and
[03:19.520 --> 03:21.780]  like, I think it's really
[03:21.780 --> 03:23.900]  easy to see how does it affect
[03:23.900 --> 03:25.660]  the world right now.
[03:26.720 --> 03:27.900]  And also
[03:27.900 --> 03:29.940]  how I learned in a hard way
[03:29.940 --> 03:31.820]  biology is way more
[03:31.820 --> 03:33.840]  complex than any malware that I ever did
[03:33.980 --> 03:35.840]  a reverse engineering in. This is
[03:35.840 --> 03:37.740]  actually one...
[03:37.740 --> 03:39.840]  just one slide of my
[03:39.840 --> 03:41.980]  class. What you see here is
[03:41.980 --> 03:44.000]  the complex system
[03:44.000 --> 03:45.880]  of what happens inside
[03:46.100 --> 03:47.720]  a cell when there is one
[03:47.720 --> 03:49.740]  ligand that is binding
[03:49.740 --> 03:51.360]  one of the receptors.
[03:52.420 --> 03:53.700]  So, apparently
[03:53.700 --> 03:56.300]  it wasn't the best idea for me.
[03:56.560 --> 03:58.300]  I'm joking, of course.
[03:58.360 --> 03:59.940]  I'm very curious and
[03:59.940 --> 04:01.720]  this area is very challenging
[04:01.720 --> 04:03.320]  for me.
[04:03.760 --> 04:05.960]  And I think
[04:05.960 --> 04:07.640]  that what we do here is just
[04:07.640 --> 04:08.680]  beautiful.
[04:10.860 --> 04:11.880]  And also
[04:11.880 --> 04:13.760]  one more thing that I want to
[04:13.760 --> 04:15.740]  point out during my lecture is that
[04:15.740 --> 04:17.660]  when I tell people about biology
[04:17.660 --> 04:19.060]  they all think that I'm
[04:19.840 --> 04:21.100]  palpating and growing
[04:22.160 --> 04:23.740]  microbes on a plate
[04:23.740 --> 04:25.020]  but actually these days
[04:26.120 --> 04:27.540]  computational biology is
[04:27.540 --> 04:29.040]  on the rise and what we are doing
[04:29.680 --> 04:32.040]  especially in Weizmann
[04:32.040 --> 04:33.740]  is trying to take
[04:33.740 --> 04:35.800]  each of the biology systems that we have
[04:35.800 --> 04:37.760]  in our body or in bacteria or anything else
[04:37.760 --> 04:39.580]  and build a model about it. This is
[04:39.580 --> 04:41.700]  also one of my classes and this is from
[04:41.700 --> 04:43.720]  Uri Alonlav that we are
[04:43.720 --> 04:45.420]  doing a collaboration with.
[04:45.420 --> 04:48.200]  So what he did here, he took
[04:48.200 --> 04:49.200]  the aldenal gland
[04:49.200 --> 04:51.520]  and tried to build
[04:51.520 --> 04:53.080]  actual equations
[04:53.080 --> 04:55.440]  I think I did something
[04:55.440 --> 04:57.700]  okay, and tried to build
[04:57.700 --> 04:59.320]  equations in order to
[04:59.320 --> 05:00.740]  understand that this
[05:01.480 --> 05:02.900]  thing here is the
[05:04.180 --> 05:05.620]  complex system of
[05:05.620 --> 05:07.580]  the glands and this one
[05:07.580 --> 05:09.760]  is some
[05:09.760 --> 05:11.580]  equations that are
[05:11.580 --> 05:13.520]  explaining why
[05:13.520 --> 05:15.180]  does diabetes type 2
[05:16.080 --> 05:16.600]  develop
[05:17.760 --> 05:20.020]  in so many people these days.
[05:20.800 --> 05:21.680]  And now I will pass
[05:21.680 --> 05:23.400]  the microphone to my friend that will
[05:23.400 --> 05:24.900]  explain to you in detail about
[05:25.400 --> 05:27.540]  some of the basic creatures that all of us
[05:27.540 --> 05:28.720]  have in our body.
[05:29.880 --> 05:32.140]  Thank you very much indeed Noor.
[05:32.760 --> 05:33.760]  By the way, I just
[05:33.760 --> 05:35.180]  want to point out that
[05:35.180 --> 05:36.860]  the person
[05:37.320 --> 05:39.060]  that actually introduced me
[05:39.060 --> 05:41.060]  into this world and made me
[05:41.060 --> 05:42.660]  very stressed
[05:44.600 --> 05:45.400]  also a very
[05:45.400 --> 05:47.300]  happy student but also a very stressed and a very
[05:47.300 --> 05:49.420]  poor student is this guy.
[05:49.420 --> 05:51.220]  So when your mom tells you not to talk
[05:51.220 --> 05:53.360]  to strangers, maybe she has a point.
[05:54.440 --> 05:55.280]  Okay, so I
[05:55.280 --> 05:57.040]  guess I don't need any further ado.
[05:57.040 --> 05:59.080]  Well, if you can see the camera, you will see that
[05:59.080 --> 06:01.320]  in contrast to Noor, I will be sitting. I'm a little older
[06:01.320 --> 06:03.780]  so forgive me for that.
[06:04.680 --> 06:06.040]  So actually
[06:06.040 --> 06:08.080]  when we were preparing this lecture
[06:09.120 --> 06:09.800]  we
[06:10.560 --> 06:11.240]  we
[06:14.520 --> 06:16.060]  we came up
[06:16.060 --> 06:18.180]  first with this title
[06:18.180 --> 06:19.680]  Back to the Future
[06:19.680 --> 06:21.660]  and me being this
[06:21.660 --> 06:24.080]  very modest, humble
[06:24.080 --> 06:26.060]  German dude that I am. I didn't
[06:26.060 --> 06:28.000]  have all these fancy animations
[06:28.000 --> 06:30.020]  that Noor showed to you.
[06:30.020 --> 06:32.000]  But actually when I envisioned this presentation
[06:32.000 --> 06:33.700]  I started with something
[06:33.700 --> 06:35.720]  lame looking like this.
[06:35.720 --> 06:37.980]  So again,
[06:37.980 --> 06:39.620]  we are from the Advisory Institute of Science
[06:39.620 --> 06:41.860]  in Israel and
[06:43.080 --> 06:43.760]  what
[06:43.760 --> 06:45.940]  I want to show you now
[06:45.940 --> 06:48.280]  in the next 25 minutes
[06:48.280 --> 06:49.760]  is in essence
[06:49.760 --> 06:51.880]  how computer science
[06:51.880 --> 06:53.860]  and biology not only go
[06:53.860 --> 06:56.240]  side by side exhibiting stunning parallels
[06:56.240 --> 06:57.860]  but also that they in fact
[06:57.860 --> 06:59.860]  have an intersection that we can
[06:59.860 --> 07:02.020]  exploit nowadays to address
[07:02.020 --> 07:03.960]  current challenges that we
[07:03.960 --> 07:05.380]  are facing.
[07:05.460 --> 07:07.540]  But before I go into this
[07:07.540 --> 07:09.960]  I actually want to fulfill
[07:09.960 --> 07:11.520]  another request that Noor
[07:11.520 --> 07:13.960]  kindly asked me which is
[07:13.960 --> 07:15.720]  to explain a very central
[07:15.720 --> 07:17.880]  concept in biology which by
[07:17.880 --> 07:19.420]  itself also has
[07:21.540 --> 07:21.940]  a counterpart
[07:21.940 --> 07:23.260]  in computer science
[07:23.690 --> 07:26.800]  and I'm talking about genetic code.
[07:26.800 --> 07:27.620]  Okay?
[07:27.620 --> 07:29.260]  Genetic code, our DNA
[07:29.260 --> 07:31.660]  is basically the language
[07:31.660 --> 07:33.760]  in which life is written.
[07:33.760 --> 07:35.100]  So that's sort of
[07:35.420 --> 07:37.720]  these chemical molecules that make up
[07:37.720 --> 07:38.840]  four letters
[07:39.500 --> 07:41.300]  A, T, C and G.
[07:41.300 --> 07:43.840]  They write a
[07:44.320 --> 07:45.660]  blueprint of life
[07:46.920 --> 07:47.480]  but as such
[07:47.480 --> 07:49.840]  they are not very useful because they have to be
[07:49.840 --> 07:52.040]  transcribed. They have to be transcribed
[07:52.040 --> 07:54.260]  from this DNA form, this genetic
[07:54.260 --> 07:55.100]  information into just another
[07:55.100 --> 07:56.860]  sort of code
[07:56.860 --> 07:59.300]  which is RNA which then
[07:59.300 --> 08:01.580]  is sort of translated
[08:01.580 --> 08:02.700]  so put into
[08:03.660 --> 08:05.040]  proteins, protein
[08:05.040 --> 08:07.200]  molecules that make up all our cells
[08:07.200 --> 08:08.560]  and therefore our body
[08:09.100 --> 08:11.040]  as a central component.
[08:11.420 --> 08:13.290]  And speaking about code, I think
[08:13.740 --> 08:15.280]  again, it's not very far
[08:15.280 --> 08:16.960]  fetched to reach from
[08:16.960 --> 08:19.040]  biology, so the science
[08:19.040 --> 08:20.440]  of life
[08:21.100 --> 08:23.140]  to computer science because
[08:23.140 --> 08:25.020]  you as computer scientists, mostly
[08:25.020 --> 08:26.760]  out there in the audience, you will know what
[08:27.360 --> 08:28.320]  a good computer
[08:29.120 --> 08:30.460]  code is.
[08:31.860 --> 08:33.000]  And basically I want to start
[08:33.000 --> 08:35.000]  by asking the question whether
[08:35.000 --> 08:37.320]  there is another analogy between
[08:37.320 --> 08:38.880]  computer science and
[08:38.880 --> 08:41.100]  biology. Are there any
[08:41.100 --> 08:43.120]  parallels in terms of computer
[08:43.120 --> 08:45.300]  viruses and biological
[08:45.300 --> 08:47.220]  viruses because, I mean,
[08:47.220 --> 08:49.280]  as you know, currently we are facing
[08:49.500 --> 08:49.920]  a global
[08:50.680 --> 08:51.880]  COVID-19
[08:51.880 --> 08:53.740]  pandemic that originates from the
[08:53.740 --> 08:55.940]  deadly spread of the
[08:55.940 --> 08:57.780]  coronavirus, SARS-CoV-2
[08:58.560 --> 08:59.840]  and I
[08:59.840 --> 09:01.980]  want to raise the question whether
[09:01.980 --> 09:04.000]  there's anything, if you
[09:04.000 --> 09:06.080]  only know computer viruses because you're
[09:06.460 --> 09:07.980]  a cyber expert, that you can
[09:07.980 --> 09:10.060]  also learn about biological
[09:10.060 --> 09:12.060]  viruses such as this
[09:12.060 --> 09:13.460]  coronavirus.
[09:13.740 --> 09:15.940]  So let's start with a computer virus, right?
[09:15.940 --> 09:17.920]  With a computer virus that you probably heard about
[09:17.920 --> 09:20.080]  the Melissa back in 1999
[09:20.080 --> 09:20.560]  which
[09:22.760 --> 09:23.480]  based on
[09:24.180 --> 09:25.900]  a sequence like a stretch of code that you
[09:25.900 --> 09:27.120]  can see on the left here
[09:28.740 --> 09:29.960]  which basically with
[09:29.960 --> 09:32.020]  some letters just gives
[09:32.020 --> 09:34.300]  also the blueprint of a
[09:34.300 --> 09:36.160]  malware basically that causes
[09:36.160 --> 09:38.080]  a lot of damage and
[09:38.080 --> 09:40.100]  I'm not sure whether you're surprised or not but what you're
[09:40.100 --> 09:42.120]  going to see on the right now here
[09:42.120 --> 09:46.080]  are the first 3,569
[09:43.920 --> 09:45.920]  letters
[09:46.080 --> 09:47.900]  of the SARS-CoV-2
[09:47.900 --> 09:50.220]  this coronavirus, the version
[09:50.220 --> 09:52.020]  Wuhan-U1
[09:52.020 --> 09:53.960]  the genome, the genetic
[09:53.960 --> 09:56.100]  information of this
[09:56.100 --> 09:57.820]  virus which in total is about
[09:57.820 --> 09:59.420]  30,000 letters long
[09:59.880 --> 10:01.960]  that makes up all
[10:01.960 --> 10:04.040]  this virus that now
[10:04.040 --> 10:05.940]  spreads on the face of the Earth
[10:05.940 --> 10:07.960]  and I'm
[10:07.960 --> 10:10.140]  saying genetic information here but to be
[10:10.140 --> 10:11.500]  precise this is not
[10:12.080 --> 10:14.100]  DNA that you're seeing but
[10:14.100 --> 10:15.680]  actually it's an RNA virus
[10:15.680 --> 10:18.100]  so while this is
[10:18.100 --> 10:20.040]  sort of the genetic information of
[10:20.040 --> 10:21.440]  this virus, it's not
[10:22.320 --> 10:24.160]  the same genetic information
[10:24.160 --> 10:26.160]  that we have in our body sense but it's
[10:26.160 --> 10:27.640]  only this transcript that I
[10:27.640 --> 10:29.700]  introduced to you before.
[10:30.080 --> 10:31.800]  But looking at this
[10:31.800 --> 10:33.960]  sequence or this source code
[10:33.960 --> 10:36.000]  of the virus, you wouldn't be
[10:36.000 --> 10:37.920]  able to judge that this is particularly harmful.
[10:37.920 --> 10:39.860]  Maybe on the left you could,
[10:39.860 --> 10:42.000]  Melissa, because you're
[10:42.000 --> 10:44.000]  an expert and you see what
[10:44.000 --> 10:44.960]  it actually does
[10:46.420 --> 10:48.000]  but there is no chance
[10:48.000 --> 10:49.780]  even the best molecular
[10:49.780 --> 10:51.840]  biologist looking at
[10:51.840 --> 10:53.920]  the code on the right would be able to
[10:53.920 --> 10:55.960]  tell, okay, this is a deadly virus.
[10:56.740 --> 10:57.780]  But still, we
[10:57.780 --> 10:59.860]  desperately needed to know this sequence
[10:59.860 --> 11:01.800]  to detect this virus
[11:02.720 --> 11:03.860]  but then the question
[11:03.860 --> 11:05.980]  is always having this sequence, having
[11:05.980 --> 11:08.000]  this source code, what does it do?
[11:08.000 --> 11:10.000]  What is this sequence translated
[11:10.000 --> 11:12.100]  to? And starting again
[11:12.100 --> 11:13.120]  with the example
[11:14.000 --> 11:16.020]  of Melissa,
[11:16.020 --> 11:17.900]  basically what this virus did
[11:17.900 --> 11:19.700]  is it prompted
[11:19.700 --> 11:21.320]  always to send
[11:22.000 --> 11:23.680]  emails to your own
[11:23.680 --> 11:25.940]  Microsoft Outlook inbox or
[11:25.940 --> 11:27.200]  started new
[11:28.000 --> 11:30.000]  Word files which basically then
[11:30.800 --> 11:31.600]  overloaded
[11:31.600 --> 11:33.260]  and completely flooded
[11:34.000 --> 11:35.940]  servers, like online servers and
[11:35.940 --> 11:37.520]  stuff. So you had,
[11:38.020 --> 11:39.560]  of course, I only made this up like
[11:39.560 --> 11:40.920]  an infinite inbox
[11:43.100 --> 11:43.540]  and
[11:43.540 --> 11:45.320]  this is sort of how
[11:45.320 --> 11:47.800]  the source code of the computer virus
[11:47.800 --> 11:49.040]  translates into
[11:49.900 --> 11:51.460]  a 3D structure, if you want to say so,
[11:51.460 --> 11:53.960]  with a malfunction, so with a bad function.
[11:54.540 --> 11:55.520]  And actually the same thing
[11:55.520 --> 11:57.520]  biologists did then, they asked the question
[11:57.520 --> 11:59.760]  okay, given the sequence of the
[11:59.760 --> 12:01.580]  coronavirus, how does this
[12:01.580 --> 12:03.540]  translate into what I mentioned before
[12:03.540 --> 12:05.520]  these proteins, right, that make
[12:05.520 --> 12:07.520]  up the shell of the
[12:07.520 --> 12:08.560]  virus and
[12:09.440 --> 12:11.260]  these different parts of the virus that
[12:11.260 --> 12:13.340]  help the virus to infect
[12:13.340 --> 12:15.680]  our body. And this is
[12:15.680 --> 12:17.540]  basically one part of this
[12:17.540 --> 12:19.360]  virus, so it's not the entire virus, it's just
[12:19.360 --> 12:21.500]  one tool, actually what I said
[12:21.500 --> 12:23.300]  like one element, one
[12:23.300 --> 12:25.520]  part of a molecule that helps this virus
[12:25.520 --> 12:27.460]  to infect our body, our
[12:27.460 --> 12:28.640]  body cells.
[12:29.380 --> 12:31.400]  And this is how this virus does
[12:31.400 --> 12:33.460]  harm, right, it infects cells and then
[12:33.460 --> 12:35.200]  eventually kills them.
[12:35.900 --> 12:37.320]  So this is how in biology
[12:37.640 --> 12:39.100]  a sequence or a source code
[12:39.100 --> 12:41.520]  translates to a structure
[12:41.520 --> 12:43.520]  and then a certain function.
[12:44.420 --> 12:45.420]  And then to
[12:45.420 --> 12:47.440]  sort of visualize how it looks
[12:47.440 --> 12:49.280]  like in total, apart from
[12:49.280 --> 12:51.400]  what it does, the
[12:51.400 --> 12:53.640]  Melissa virus to overload
[12:53.640 --> 12:55.240]  your inbox and the coronavirus
[12:55.240 --> 12:58.760]  to infect your body cells,
[12:58.760 --> 12:59.460]  actually
[12:59.460 --> 13:01.360]  for the Melissa virus then it took
[13:01.360 --> 13:03.480]  an artist to visualize
[13:03.480 --> 13:05.340]  how this overload looks like
[13:05.340 --> 13:07.400]  and this is how an artist apparently pictures
[13:07.400 --> 13:09.160]  how a computer virus
[13:09.160 --> 13:11.440]  looks like, which as you
[13:11.440 --> 13:13.340]  can probably better judge than I do
[13:14.180 --> 13:15.220]  is just
[13:15.220 --> 13:17.260]  more or less arbitrary
[13:17.260 --> 13:18.920]  representation.
[13:19.440 --> 13:21.400]  But in fact, in biology, again, we can
[13:21.400 --> 13:23.360]  use this wonderful
[13:23.360 --> 13:25.340]  invention of microscopes to
[13:25.340 --> 13:27.500]  actually look at how the virus looks
[13:27.500 --> 13:29.660]  like. However, conventional
[13:29.660 --> 13:31.620]  light microscopes are not sufficient enough
[13:31.620 --> 13:33.760]  here, so actually we need electron
[13:33.760 --> 13:35.740]  microscopes. However, if we use
[13:35.740 --> 13:37.400]  these very sophisticated
[13:38.020 --> 13:39.680]  apparatuses that cost like
[13:39.840 --> 13:41.600]  a million dollars each and
[13:41.600 --> 13:43.780]  require a lot of expertise to use them,
[13:43.780 --> 13:45.980]  then we can actually see a virus inside
[13:45.980 --> 13:47.880]  of cells.
[13:47.880 --> 13:49.440]  So all these blue particles
[13:49.440 --> 13:51.880]  that in this Forbes article
[13:51.880 --> 13:53.100]  are described as
[13:53.840 --> 13:55.900]  pepperoni, sorry,
[13:55.900 --> 13:58.020]  slices of sausage
[13:58.020 --> 14:00.000]  on your pizza, they are
[14:00.000 --> 14:02.160]  the actual virus. However, I also must
[14:02.160 --> 14:03.540]  say the truth is
[14:03.540 --> 14:06.140]  if you use an electron microscope to observe
[14:06.140 --> 14:07.000]  the coronavirus
[14:08.840 --> 14:10.500]  in cells,
[14:10.500 --> 14:11.960]  so this is actually the coronavirus of
[14:11.960 --> 14:13.920]  the first identified
[14:13.920 --> 14:15.120]  US patient
[14:15.120 --> 14:18.240]  that was imaged here,
[14:18.240 --> 14:20.020]  this original picture from
[14:20.020 --> 14:22.480]  the electron microscope had no color.
[14:22.720 --> 14:24.060]  So this blue
[14:24.060 --> 14:25.620]  color is just put up
[14:25.620 --> 14:27.860]  artificially in the end to visualize
[14:27.860 --> 14:29.700]  the virus particles.
[14:30.980 --> 14:31.720]  But here
[14:31.720 --> 14:33.000]  we are still talking about
[14:34.240 --> 14:34.760]  only
[14:36.560 --> 14:37.620]  the size of
[14:38.300 --> 14:39.720]  a meter divided by
[14:39.720 --> 14:41.600]  one billion. So it's
[14:41.600 --> 14:43.400]  in this order of magnitude.
[14:43.400 --> 14:45.080]  So it's very small.
[14:46.300 --> 14:47.540]  And you need sophisticated,
[14:47.540 --> 14:49.420]  as I said, microscopes to actually see it,
[14:49.420 --> 14:51.440]  but you can see it. You can understand
[14:51.440 --> 14:53.460]  how this sequence source
[14:53.460 --> 14:55.500]  code information that we saw,
[14:55.500 --> 14:57.140]  how this translates to
[14:57.140 --> 14:59.140]  first individual
[14:59.140 --> 15:01.340]  aspects of this virus,
[15:01.340 --> 15:04.020]  individual molecules, and then the virus as such.
[15:04.020 --> 15:05.580]  Now, once it
[15:05.580 --> 15:07.620]  has done its terrible job infecting
[15:07.620 --> 15:09.300]  our body cells.
[15:09.640 --> 15:11.580]  And this sort of
[15:12.400 --> 15:13.480]  shows that
[15:13.480 --> 15:15.200]  to some extent there are
[15:15.200 --> 15:17.360]  parallels between computer science and
[15:17.360 --> 15:19.320]  biology, in this case computer
[15:19.320 --> 15:20.820]  viruses and
[15:20.820 --> 15:22.780]  biological viruses such as
[15:22.780 --> 15:24.540]  SARS-CoV-2.
[15:26.200 --> 15:26.540]  But
[15:26.540 --> 15:28.640]  is there, you know, besides
[15:29.980 --> 15:30.540]  from
[15:32.040 --> 15:33.400]  this example,
[15:33.400 --> 15:34.960]  is there something we can
[15:34.960 --> 15:37.040]  do about this? Is there something we can use
[15:37.040 --> 15:38.640]  from computer science
[15:38.640 --> 15:40.860]  to help us
[15:40.860 --> 15:42.360]  fighting this
[15:42.360 --> 15:44.920]  terrible pandemic that we are facing
[15:44.920 --> 15:47.120]  in the US and all over the globe?
[15:47.220 --> 15:48.880]  Is there an intersection between computer
[15:48.880 --> 15:50.580]  science and biological aspects of
[15:50.580 --> 15:52.320]  this pandemic?
[15:52.420 --> 15:54.760]  This is a question Noah and I basically
[15:54.760 --> 15:56.960]  asked ourselves. And in the very
[15:56.960 --> 15:58.220]  beginning of this pandemic
[15:58.700 --> 16:00.740]  in March, in fact, we
[16:00.740 --> 16:02.860]  participated in
[16:03.700 --> 16:04.840]  a hackathon that was
[16:04.840 --> 16:06.780]  initiated by
[16:06.780 --> 16:08.780]  InDev in collaboration
[16:08.780 --> 16:10.560]  with Microsoft. They called it
[16:10.560 --> 16:12.000]  HackCorona.
[16:12.440 --> 16:15.180]  Creativity will not be quarantined.
[16:15.180 --> 16:16.160]  And in the next
[16:16.160 --> 16:18.520]  five to ten minutes, I would like
[16:18.520 --> 16:20.540]  to introduce to you this
[16:20.540 --> 16:22.500]  project that we initiated there to show
[16:22.500 --> 16:24.620]  you how computer science can actually facilitate
[16:24.620 --> 16:26.280]  in a very
[16:26.280 --> 16:28.420]  concrete example, our
[16:29.280 --> 16:30.000]  fight
[16:30.000 --> 16:32.440]  against this virus.
[16:33.340 --> 16:34.080]  And the
[16:35.400 --> 16:36.420]  question that
[16:36.420 --> 16:38.240]  we asked basically is
[16:39.800 --> 16:40.520]  how
[16:40.520 --> 16:42.400]  can we use computers to
[16:42.400 --> 16:44.280]  combat COVID-19?
[16:44.280 --> 16:46.220]  And the particular question we were wondering
[16:46.220 --> 16:48.280]  about, because this was of major
[16:48.280 --> 16:50.420]  concern in the beginning of this pandemic,
[16:50.420 --> 16:52.080]  is how to
[16:52.080 --> 16:54.180]  help people who develop severe
[16:54.180 --> 16:56.060]  complications. Because we
[16:56.060 --> 16:58.800]  understood that asymptomatic patients,
[16:58.800 --> 17:00.320]  of course, they are very dangerous in
[17:00.320 --> 17:02.180]  terms of unobserved
[17:02.180 --> 17:04.580]  spreaders of this disease.
[17:04.820 --> 17:06.380]  But in case
[17:06.380 --> 17:08.340]  someone develops severe complications,
[17:08.340 --> 17:11.120]  needs ventilation in the hospital,
[17:11.120 --> 17:12.380]  then this
[17:12.380 --> 17:13.040]  might
[17:14.820 --> 17:16.520]  present a major challenge for
[17:16.520 --> 17:18.320]  the health system in Israel
[17:18.320 --> 17:20.040]  and all over the world.
[17:20.680 --> 17:22.560]  So how can we basically assess
[17:22.560 --> 17:24.700]  the uncertainty whether a certain
[17:24.700 --> 17:26.480]  COVID-19 patient will develop
[17:26.480 --> 17:27.340]  severe
[17:28.520 --> 17:29.820]  symptoms and require
[17:31.160 --> 17:32.650]  hospitalization and ventilation?
[17:33.820 --> 17:34.500]  This is
[17:34.500 --> 17:36.380]  the question we asked. And as you can see
[17:36.380 --> 17:38.280]  in this picture, again, we were
[17:38.280 --> 17:40.440]  pretty much using
[17:40.440 --> 17:42.540]  biology as well as computer science
[17:42.540 --> 17:44.280]  by these two mice that you can see here,
[17:44.280 --> 17:46.400]  like a computer mouse and an actual mouse,
[17:46.400 --> 17:48.860]  or at least a stuffed animal of a mouse.
[17:50.540 --> 17:52.320]  And so, in simple
[17:52.320 --> 17:54.740]  terms, our question was,
[17:54.740 --> 17:56.260]  given the fact that we have
[17:56.420 --> 17:58.040]  a certain number of
[17:58.040 --> 17:59.580]  COVID-19 patients
[18:00.640 --> 18:02.100]  that require, because
[18:02.100 --> 18:04.580]  they develop severe complications,
[18:04.580 --> 18:06.060]  hospital beds,
[18:06.060 --> 18:08.140]  how to deal with them if the number of
[18:08.140 --> 18:10.100]  hospital beds is actually
[18:10.100 --> 18:12.120]  lower than the number of patients who require
[18:12.120 --> 18:14.240]  hospitalization? In other
[18:14.240 --> 18:16.600]  words, five
[18:16.600 --> 18:18.220]  exceeds three, right?
[18:18.220 --> 18:19.780]  So how to decide
[18:19.780 --> 18:21.920]  which of them to allocate
[18:21.920 --> 18:23.880]  to the hospital?
[18:23.880 --> 18:25.980]  How to estimate what's the fraction of
[18:25.980 --> 18:28.100]  people who develop severe complications
[18:28.100 --> 18:30.120]  and how to decide whom to offer
[18:30.120 --> 18:32.000]  with a certain priority these
[18:32.000 --> 18:33.880]  available resources?
[18:33.880 --> 18:36.060]  This was the question.
[18:36.060 --> 18:37.560]  And the reason why we deal with this
[18:37.560 --> 18:40.140]  was because at the Weizmann Institute
[18:40.140 --> 18:41.880]  we are really at the forefront
[18:41.880 --> 18:43.880]  of this, not only that we have
[18:43.880 --> 18:45.880]  medical doctors on campus
[18:45.880 --> 18:47.880]  who see these patients on a day-to-day
[18:47.880 --> 18:49.920]  basis, but also the labs
[18:49.920 --> 18:51.580]  we're working in from
[18:51.580 --> 18:54.080]  Prof. Dr. Dr. Eran
[18:54.080 --> 18:56.360]  Elinav and Prof. Dr. Ido Amit,
[18:56.360 --> 18:57.780]  they develop now
[18:57.780 --> 19:00.080]  technologies to not only
[19:02.380 --> 19:03.500]  detect these
[19:03.500 --> 19:05.340]  infections and stratify
[19:05.340 --> 19:07.460]  these patients, but also streamline this
[19:07.460 --> 19:09.120]  process, really linking
[19:09.120 --> 19:11.280]  the foundational research with
[19:11.280 --> 19:13.760]  the application in the clinics.
[19:14.100 --> 19:14.780]  Okay?
[19:14.780 --> 19:17.640]  So this is why we also felt
[19:17.640 --> 19:19.540]  dedicated to address this
[19:19.540 --> 19:21.620]  with our expertise, which
[19:21.620 --> 19:22.960]  comes more from the
[19:23.500 --> 19:25.240]  computational side here.
[19:25.240 --> 19:27.540]  So what we did in the beginning
[19:27.540 --> 19:29.500]  is, first of all, we asked, okay,
[19:29.500 --> 19:31.780]  what's the data we are dealing with?
[19:31.780 --> 19:33.720]  Because only if you have data,
[19:33.720 --> 19:35.800]  you can actually use computational routines
[19:35.800 --> 19:37.980]  to address
[19:37.980 --> 19:39.980]  all these issues.
[19:40.640 --> 19:41.660]  So we
[19:41.660 --> 19:43.680]  basically simulated
[19:43.680 --> 19:45.920]  based on existing
[19:45.920 --> 19:47.780]  data,
[19:47.780 --> 19:48.640]  an Israeli
[19:49.780 --> 19:51.840]  dataset comprised of 10,000
[19:51.840 --> 19:53.980]  cases, where we really used
[19:53.980 --> 19:55.100]  Israeli demographic
[19:55.780 --> 19:57.860]  information and health record data
[19:57.860 --> 19:59.400]  that was publicly available
[19:59.400 --> 20:01.760]  to a certain extent,
[20:01.760 --> 20:03.660]  and set up
[20:03.660 --> 20:05.420]  sort of a database where
[20:05.420 --> 20:07.240]  we can get an idea, okay, this is Israeli
[20:07.780 --> 20:09.240]  population, and now
[20:09.240 --> 20:10.220]  given all this
[20:11.680 --> 20:13.700]  demographic information and health record information,
[20:13.700 --> 20:14.860]  can we basically now
[20:15.680 --> 20:17.360]  estimate what's the fraction of people who
[20:17.360 --> 20:19.620]  might develop severe COVID-19
[20:19.620 --> 20:21.180]  complications? So first of all,
[20:21.180 --> 20:23.360]  as you can see, we are struggling entering all this
[20:23.360 --> 20:25.280]  data, but then what we obtained
[20:25.280 --> 20:27.100]  is basically a more or less realistic
[20:27.100 --> 20:29.160]  Israeli age distribution
[20:29.160 --> 20:31.340]  here, and
[20:31.340 --> 20:33.160]  also what seems to be
[20:33.160 --> 20:34.840]  important, and we will hear about
[20:34.840 --> 20:37.180]  diabetes and the metabolic syndrome later
[20:37.180 --> 20:38.740]  on in Noah's part,
[20:38.740 --> 20:41.340]  the distribution of the body mass
[20:41.340 --> 20:42.960]  index in the Israeli
[20:42.960 --> 20:44.580]  population,
[20:45.100 --> 20:47.420]  which is basically a measure for your
[20:47.420 --> 20:49.360]  body density,
[20:49.360 --> 20:51.260]  so it's your body size divided
[20:51.260 --> 20:52.340]  by your
[20:53.080 --> 20:55.280]  body height to the
[20:55.280 --> 20:57.300]  square, and it gives you an estimate
[20:58.760 --> 20:59.160]  of...
[20:59.640 --> 21:01.360]  if you have a high BMI,
[21:01.360 --> 21:03.180]  you are considered obese, and then actually
[21:03.180 --> 21:05.040]  this also affects your risk of
[21:05.040 --> 21:07.340]  COVID-19 severe
[21:07.340 --> 21:09.080]  complications, but also whether you get infected
[21:09.080 --> 21:11.260]  in the first place. But it's not fully
[21:11.260 --> 21:13.060]  understood, it's just an observation, so
[21:13.060 --> 21:15.300]  anyways, we thought we need to include
[21:15.300 --> 21:17.400]  this information, so first of all, we needed to have
[21:17.400 --> 21:19.360]  some sort of realistic distribution
[21:19.960 --> 21:21.240]  of this kind of
[21:21.240 --> 21:23.320]  data. And then
[21:23.320 --> 21:25.520]  what we were basically sampling from this information
[21:25.520 --> 21:27.100]  we had back then when Israel
[21:27.100 --> 21:29.440]  was not really facing many
[21:29.440 --> 21:31.240]  many cases at the time
[21:31.240 --> 21:33.180]  was this sort of distribution
[21:33.180 --> 21:35.240]  of people
[21:35.240 --> 21:37.420]  who got infected and then
[21:37.420 --> 21:38.700]  sort of were
[21:39.320 --> 21:41.340]  surviving through the disease, and
[21:41.340 --> 21:43.380]  people who eventually died, and then you can
[21:43.380 --> 21:45.300]  see that even though the overall frequency
[21:45.300 --> 21:47.460]  of dead people, of course, or of dying
[21:47.460 --> 21:49.320]  people is with the
[21:49.320 --> 21:51.260]  infection or upon infection,
[21:51.260 --> 21:53.560]  not necessarily caused by the infection,
[21:53.560 --> 21:55.340]  is very low, but you can see that
[21:56.880 --> 21:58.720]  if you're older, the
[21:59.260 --> 22:01.300]  probability that you actually die
[22:01.300 --> 22:03.680]  after getting infected with SARS-CoV-2
[22:03.680 --> 22:05.660]  is higher.
[22:07.180 --> 22:08.720]  So what we then
[22:09.140 --> 22:11.260]  did is we used this data set,
[22:11.260 --> 22:13.060]  this information that we
[22:13.060 --> 22:15.220]  compiled, and
[22:15.220 --> 22:17.580]  trained the model to basically
[22:17.580 --> 22:19.180]  predict, given
[22:19.520 --> 22:20.640]  a certain patient with
[22:21.100 --> 22:23.040]  this health record and demographic
[22:23.040 --> 22:24.740]  information,
[22:24.740 --> 22:26.880]  what's the risk of developing
[22:26.880 --> 22:29.220]  severe complications, and
[22:29.220 --> 22:31.200]  then what you can always do is you can compare
[22:31.200 --> 22:33.280]  your prediction to the actually observed
[22:33.280 --> 22:35.120]  risk, because you have
[22:35.120 --> 22:36.660]  information on this, right?
[22:36.820 --> 22:39.080]  And then we get a pretty good agreement between
[22:39.080 --> 22:40.960]  the predicted risk and the observed
[22:40.960 --> 22:42.340]  risk here in this
[22:43.040 --> 22:44.780]  preliminary data set.
[22:45.100 --> 22:47.300]  And then we can ask the question, okay, what kind of information
[22:47.300 --> 22:48.860]  is actually necessary to
[22:49.700 --> 22:51.620]  predict this risk in a
[22:51.620 --> 22:53.460]  reliable manner? And then
[22:53.880 --> 22:55.860]  just as a test case,
[22:55.860 --> 22:57.640]  proof of principle, we said,
[22:57.640 --> 23:00.100]  okay, age seems to be
[23:00.100 --> 23:01.420]  relevant and
[23:02.820 --> 23:03.680]  critical,
[23:03.680 --> 23:05.320]  so what happens if
[23:06.600 --> 23:07.460]  we are
[23:07.460 --> 23:09.400]  leaving out the age information? And what you
[23:09.400 --> 23:11.620]  are seeing here is basically a receiver
[23:11.620 --> 23:13.940]  operating characteristic curve,
[23:13.940 --> 23:15.160]  a rock plot,
[23:15.160 --> 23:17.120]  and basically if your
[23:17.120 --> 23:19.620]  line, so your specificity
[23:19.620 --> 23:20.280]  and sensitivity
[23:21.240 --> 23:23.440]  of this prediction performs
[23:23.440 --> 23:25.220]  pretty much randomly,
[23:25.220 --> 23:27.240]  then this line that you see is close
[23:27.240 --> 23:29.320]  to the diagonal line. So as you can see,
[23:29.320 --> 23:32.020]  the orange line, if we leave out the age information,
[23:32.020 --> 23:33.220]  we cannot really predict
[23:33.220 --> 23:35.200]  that it's like randomly drawing,
[23:35.200 --> 23:37.240]  okay, this person will develop
[23:37.240 --> 23:39.340]  complications or this person
[23:39.340 --> 23:41.420]  won't, but if we just consider
[23:41.420 --> 23:43.200]  the age information as well in addition
[23:43.200 --> 23:45.140]  to all the other information, then
[23:45.820 --> 23:47.220]  this prediction becomes
[23:47.220 --> 23:49.380]  very sensitive and very specific.
[23:49.380 --> 23:51.200]  So it means that if we
[23:51.200 --> 23:53.120]  say a person does
[23:53.120 --> 23:54.800]  develop complications, then
[23:55.360 --> 23:56.980]  the chance that this is a true
[23:56.980 --> 23:58.680]  positive indication is
[23:59.900 --> 24:01.260]  high. But also
[24:01.260 --> 24:03.620]  if we say this person won't develop complications,
[24:03.620 --> 24:05.120]  then also the chances
[24:05.120 --> 24:07.160]  that this is a true negative, so it
[24:07.160 --> 24:09.300]  indeed won't develop complications,
[24:09.300 --> 24:11.180]  it's also sort of high.
[24:11.180 --> 24:12.860]  So of course the perfect prediction
[24:12.860 --> 24:15.020]  would be just going straight up
[24:15.020 --> 24:16.920]  and then straight to the right,
[24:16.920 --> 24:18.980]  but also in this toy model
[24:18.980 --> 24:21.080]  or preliminary thing,
[24:21.080 --> 24:22.600]  the performance of this prediction
[24:22.600 --> 24:22.980]  was
[24:25.160 --> 24:27.380]  not so bad, I would say.
[24:27.660 --> 24:28.700]  So the way we
[24:28.700 --> 24:30.280]  envision this basically is
[24:31.140 --> 24:32.760]  once people get sick or
[24:32.760 --> 24:35.200]  are suspected of being infected,
[24:35.200 --> 24:36.340]  they have to see a doctor
[24:36.340 --> 24:38.780]  who then puts in
[24:38.780 --> 24:40.700]  all the information into a certain
[24:40.700 --> 24:42.780]  database and some
[24:42.780 --> 24:44.060]  device that collects
[24:44.640 --> 24:46.960]  this information and then basically trains
[24:46.960 --> 24:48.880]  this model that is based
[24:48.880 --> 24:50.320]  on machine learning algorithms
[24:51.300 --> 24:52.800]  and all these health record
[24:52.800 --> 24:55.020]  information and then basically puts out
[24:55.020 --> 24:56.260]  back to the doctor
[24:57.520 --> 24:58.880]  a certain risk score.
[24:58.880 --> 25:00.620]  So the application
[25:00.620 --> 25:02.300]  tells the doctor, okay, this person
[25:02.300 --> 25:04.820]  exhibits a certain probability to develop
[25:04.820 --> 25:06.940]  severe complications, so better save a hospital
[25:06.940 --> 25:08.920]  bed for this one, or in the ideal case
[25:10.580 --> 25:12.460]  this person can just be
[25:13.300 --> 25:14.920]  isolated at home, there is
[25:15.080 --> 25:17.600]  a very low chance that it's going to be severe,
[25:17.600 --> 25:19.360]  just monitor it carefully,
[25:19.360 --> 25:20.200]  but most likely
[25:20.800 --> 25:22.800]  the hospital bed allocated to this
[25:23.660 --> 25:24.980]  patient can be
[25:24.980 --> 25:27.300]  given to somebody else with a higher risk.
[25:28.860 --> 25:30.840]  So, wait, what are we
[25:31.580 --> 25:33.100]  lacking here? I already
[25:33.100 --> 25:34.960]  presented to you that this machine learning algorithm
[25:34.960 --> 25:36.980]  sort of works, but we need
[25:36.980 --> 25:39.060]  this device, right? We need this application.
[25:39.600 --> 25:40.720]  And this is also what we
[25:40.720 --> 25:42.520]  sort of developed, like
[25:43.080 --> 25:45.100]  something that can work on your mobile phone
[25:45.100 --> 25:46.920]  and helps medical doctors, again, that
[25:46.920 --> 25:48.940]  we have in our group and we talk to them on a daily
[25:48.940 --> 25:51.020]  basis, to just use
[25:51.020 --> 25:52.940]  it in a very handy way. If they
[25:52.940 --> 25:54.320]  see the patient, they can check
[25:54.740 --> 25:57.200]  in all this information in less than a
[25:57.200 --> 25:59.140]  minute, and
[25:59.140 --> 26:00.920]  then basically we can
[26:03.220 --> 26:04.220]  just let
[26:04.220 --> 26:06.580]  the machine learning algorithm do
[26:06.580 --> 26:08.600]  its job and see whether there's
[26:09.020 --> 26:10.140]  a certain risk.
[26:10.660 --> 26:12.020]  And, of course, this
[26:12.660 --> 26:14.300]  machine learning algorithm that I presented
[26:14.300 --> 26:16.360]  to you, this model that we used there to make
[26:16.360 --> 26:18.380]  the predictions, was just the first attempt
[26:18.380 --> 26:20.240]  that we had in the Sakaton, you know, very
[26:20.240 --> 26:21.160]  coarse-grained
[26:22.120 --> 26:23.880]  first rough trial.
[26:25.320 --> 26:26.320]  Needless to
[26:26.320 --> 26:27.800]  say, you can also use
[26:27.800 --> 26:30.080]  more sophisticated methods
[26:30.080 --> 26:31.300]  where you can use all these
[26:32.480 --> 26:32.880]  different
[26:34.440 --> 26:36.100]  biological information on these patients
[26:36.100 --> 26:38.200]  to assign how
[26:38.200 --> 26:40.140]  important is this
[26:41.760 --> 26:42.160]  information
[26:42.160 --> 26:44.260]  to either say that there is a high
[26:44.260 --> 26:46.200]  risk, or maybe it's just informative to say that there's
[26:46.300 --> 26:48.200]  a low risk, or maybe a combination of both,
[26:48.200 --> 26:50.080]  right? So we can, of course, you
[26:50.080 --> 26:52.180]  always want more data. We can try
[26:52.180 --> 26:54.220]  different models, and as
[26:54.220 --> 26:56.340]  I said, we are in day-to-day contact with the clinicians
[26:56.340 --> 26:58.160]  actually in charge, actually doing this, actually
[26:58.160 --> 27:00.000]  doing the anamnesis or talking to the
[27:00.000 --> 27:01.800]  patients,
[27:01.800 --> 27:04.500]  suspected infected people.
[27:05.720 --> 27:05.860]  So
[27:05.860 --> 27:08.140]  this is what we envisioned,
[27:08.140 --> 27:09.960]  you know, we can't really
[27:09.960 --> 27:12.220]  not necessarily change
[27:12.220 --> 27:14.220]  the number of people who are
[27:14.220 --> 27:16.000]  susceptible and eventually will get
[27:16.000 --> 27:18.520]  infected. It's just a matter of time,
[27:18.520 --> 27:19.860]  it seems like.
[27:21.160 --> 27:22.200]  And maybe
[27:22.200 --> 27:23.940]  also, you know, the health system is sort of
[27:23.940 --> 27:26.020]  currently at its limit, so we cannot
[27:26.020 --> 27:28.800]  really expand the number of hospital beds.
[27:29.020 --> 27:30.120]  But this app
[27:30.120 --> 27:31.620]  that we developed, that by the way we
[27:31.620 --> 27:33.900]  called Virologens, because it's like an
[27:33.900 --> 27:36.160]  intelligent solution to virology
[27:36.160 --> 27:37.380]  and inference of
[27:38.460 --> 27:39.000]  this
[27:40.800 --> 27:41.880]  virus-induced
[27:41.880 --> 27:44.040]  health complications,
[27:44.040 --> 27:46.320]  we could now use to assign
[27:46.320 --> 27:47.900]  again hospital
[27:47.900 --> 27:50.060]  beds to people who need it the most.
[27:50.060 --> 27:51.960]  So in the scope of this hackathon, which of course is
[27:51.960 --> 27:54.320]  very preliminary, and we are still sort of
[27:55.380 --> 27:56.480]  thinking of how to
[27:56.480 --> 27:58.180]  put this further,
[27:58.180 --> 28:00.100]  in the scope of this hackathon, this
[28:00.520 --> 28:01.920]  question was solved.
[28:02.440 --> 28:04.900]  But as I already indicated, in the scope of this hackathon,
[28:04.900 --> 28:06.440]  and this is just COVID-19, which
[28:06.440 --> 28:08.380]  at the moment seems to be a major part of our
[28:08.380 --> 28:10.480]  lives, but certainly it's not the only part
[28:10.480 --> 28:12.340]  of our life, so
[28:13.860 --> 28:14.600]  can we
[28:15.020 --> 28:16.340]  make a more general
[28:16.940 --> 28:18.340]  observation when we look at
[28:18.340 --> 28:20.700]  computer science and biology?
[28:22.040 --> 28:22.940]  Or in
[28:22.940 --> 28:24.800]  other words, how can we
[28:25.600 --> 28:26.940]  combine powerful
[28:26.940 --> 28:28.960]  algorithms from computer science
[28:28.960 --> 28:31.200]  with biological data sets?
[28:31.200 --> 28:32.860]  And this is what I've shown,
[28:32.860 --> 28:34.660]  what I'm showing you here, it's basically
[28:34.660 --> 28:36.680]  the intersection of biological
[28:36.680 --> 28:38.780]  data, this is how biological data
[28:38.780 --> 28:41.000]  nowadays sort of looks like.
[28:42.060 --> 28:42.920]  How can
[28:42.920 --> 28:44.880]  we combine this biology with
[28:44.880 --> 28:46.740]  this more like a conceptual or
[28:46.740 --> 28:48.340]  theoretical view of
[28:48.340 --> 28:50.720]  biology as a system?
[28:51.080 --> 28:52.480]  As a living
[28:52.480 --> 28:54.620]  system that just operates
[28:54.620 --> 28:56.620]  certain commands and has
[28:56.620 --> 28:58.540]  some input-output functions.
[28:58.840 --> 29:00.480]  And the answer here is also the computer,
[29:00.480 --> 29:02.400]  because only the computer will help us to process
[29:02.400 --> 29:04.300]  this data and to put it into these
[29:04.300 --> 29:06.360]  sort of models in terms of
[29:06.360 --> 29:08.480]  understanding how biological systems
[29:08.480 --> 29:10.540]  work. And this is a discipline
[29:10.540 --> 29:12.540]  that I actually studied, I did my
[29:12.540 --> 29:14.720]  PhD in this field,
[29:14.720 --> 29:16.880]  which is called systems biology.
[29:17.760 --> 29:18.880]  And let me
[29:18.880 --> 29:20.740]  introduce to you the field of systems
[29:20.740 --> 29:22.960]  biology with one of the godfathers of systems
[29:22.960 --> 29:24.780]  biology, a guy who actually does
[29:24.780 --> 29:26.740]  also a lot of robotics and
[29:26.740 --> 29:29.020]  intelligent design and stuff, a Japanese
[29:29.020 --> 29:30.920]  guy who also invented
[29:30.920 --> 29:32.960]  robots that help people
[29:32.960 --> 29:34.720]  to rescue
[29:34.720 --> 29:36.160]  people from
[29:37.140 --> 29:38.320]  collapsed buildings
[29:39.080 --> 29:41.160]  after earthquakes and stuff.
[29:42.360 --> 29:42.960]  Hiroaki
[29:42.960 --> 29:44.820]  Kitano is his name, and
[29:44.820 --> 29:46.700]  he describes systems biology as the
[29:46.700 --> 29:48.880]  integration of experimental and
[29:48.880 --> 29:51.000]  computational approaches to precisely
[29:51.000 --> 29:52.620]  characterize biological systems
[29:52.620 --> 29:55.020]  with quantitative data.
[29:55.100 --> 29:56.260]  Okay, so now we have
[29:56.680 --> 29:59.120]  quantitative data, which is machine-readable
[29:59.120 --> 30:00.220]  from biology,
[30:00.620 --> 30:03.020]  that we will use
[30:03.020 --> 30:05.100]  to understand, in combination
[30:05.100 --> 30:07.000]  with experiments in the laboratory,
[30:07.740 --> 30:08.880]  how this
[30:08.880 --> 30:10.860]  biological system works, simulating it at the
[30:10.860 --> 30:12.860]  computer, using algorithms to
[30:12.860 --> 30:14.960]  understand what's going on
[30:14.960 --> 30:16.640]  there. And what does
[30:16.640 --> 30:19.040]  go on there? How does this look like?
[30:19.560 --> 30:20.940]  Well, you can imagine
[30:20.940 --> 30:22.840]  as this sort of simple diagram
[30:22.840 --> 30:24.700]  here, where basically on the
[30:24.700 --> 30:26.560]  x-axis, you will have
[30:26.560 --> 30:28.880]  time, and on the y-axis,
[30:28.880 --> 30:30.880]  you will have the amount, quantitative
[30:30.880 --> 30:33.000]  data, okay, of a biological
[30:33.000 --> 30:35.340]  entity. Now,
[30:36.780 --> 30:38.660]  what does this mean? So basically,
[30:38.660 --> 30:40.840]  what you have in blue, these blue dots, these are
[30:40.840 --> 30:42.860]  your data points, and then
[30:42.860 --> 30:45.020]  with the computer, you can basically simulate the behavior
[30:45.020 --> 30:46.660]  of the system, and you have some mean
[30:46.660 --> 30:49.180]  simulation here, the blue line,
[30:49.180 --> 30:50.880]  but then also you have this shaded region, which is
[30:50.880 --> 30:52.840]  basically your uncertainty, okay,
[30:52.840 --> 30:54.860]  your confidence interval. So not
[30:54.860 --> 30:57.100]  only you get an understanding with the computer,
[30:57.100 --> 30:59.020]  how does the system change over time,
[30:59.020 --> 31:00.760]  how do biological entities
[31:00.760 --> 31:02.660]  change over time,
[31:02.660 --> 31:04.900]  but also, how certain are
[31:04.900 --> 31:06.680]  you that this happens, right? The virus can
[31:06.680 --> 31:08.840]  spread, like,
[31:08.840 --> 31:10.780]  enormously, but maybe there's
[31:10.780 --> 31:12.860]  also a chance it doesn't spread, or maybe there's a chance
[31:12.860 --> 31:14.980]  that it spreads, like, even more.
[31:15.240 --> 31:16.840]  And this is something we can quantify,
[31:16.840 --> 31:18.900]  we can calculate. Sometimes
[31:18.900 --> 31:20.840]  we have a problem with communicating this, but
[31:20.840 --> 31:22.840]  there are possibilities in mathematics
[31:22.840 --> 31:24.880]  to calculate these confidence
[31:24.880 --> 31:27.120]  intervals and these uncertainties.
[31:28.120 --> 31:28.840]  The way
[31:28.840 --> 31:30.740]  this is done, the way we describe how
[31:30.880 --> 31:32.920]  a biological entity changes over time
[31:32.920 --> 31:34.800]  is, in mathematics,
[31:34.800 --> 31:36.820]  called ordinary differential
[31:36.820 --> 31:38.920]  equation. So this basically
[31:38.920 --> 31:40.940]  describes to you how your biological entity
[31:40.940 --> 31:42.900]  B is changing,
[31:42.900 --> 31:44.740]  so the D represents the change, okay, it's the
[31:44.740 --> 31:46.580]  change of B over
[31:46.580 --> 31:48.700]  the change of time T, okay?
[31:48.700 --> 31:51.000]  How does B change with time?
[31:51.200 --> 31:52.800]  And this is now depending on
[31:52.800 --> 31:54.940]  certain parameters, P, P1
[31:54.940 --> 31:56.980]  and P2. And the plus
[31:56.980 --> 31:59.040]  always means, you know, you get something more,
[31:59.040 --> 32:00.720]  so depending on this parameter,
[32:00.720 --> 32:02.840]  P1, you get more of B if you
[32:02.840 --> 32:04.840]  already have B. So
[32:04.840 --> 32:06.800]  if you have
[32:07.340 --> 32:09.920]  something that can sort of reproduce,
[32:09.920 --> 32:11.140]  like a virus can reproduce
[32:11.140 --> 32:12.880]  or we can reproduce, you know, you
[32:12.880 --> 32:15.240]  will get even more if you have more people to start
[32:15.240 --> 32:17.260]  with, right? And then basically
[32:17.260 --> 32:19.260]  the reproduction rate is your parameter
[32:19.260 --> 32:21.180]  P, and it's positive because you get
[32:21.180 --> 32:23.080]  more. But then
[32:23.080 --> 32:25.220]  this curve of B is not only going up, it's not
[32:25.220 --> 32:27.200]  only just getting more with time, right? It's also
[32:27.200 --> 32:29.180]  going down. So there's also
[32:29.180 --> 32:31.420]  another parameter, P2, and there's a minus
[32:31.420 --> 32:33.260]  because it's getting less, that
[32:33.260 --> 32:34.280]  tells me that
[32:35.600 --> 32:37.400]  B is... the change of B over time
[32:37.400 --> 32:39.340]  is getting negative. There's something
[32:40.260 --> 32:41.420]  subtracted from it.
[32:41.620 --> 32:43.220]  And of course, the more B we have, the more likely
[32:43.220 --> 32:45.260]  it is that, you know, it destructs itself. You know,
[32:45.260 --> 32:47.240]  humanity is very good at destructing itself
[32:47.240 --> 32:48.580]  somehow, unfortunately.
[32:49.560 --> 32:51.320]  But just by itself, it won't go just
[32:51.320 --> 32:52.740]  up and down there. We need
[32:53.320 --> 32:55.560]  another component here that influences this.
[32:55.560 --> 32:57.280]  And this other component is A.
[32:57.740 --> 32:59.320]  So for A, we can actually
[32:59.320 --> 33:01.460]  write down a very similar equation.
[33:01.460 --> 33:03.300]  So how does A change over time? Well,
[33:03.300 --> 33:05.460]  it grows, but not only depending on A
[33:05.460 --> 33:07.400]  but also on B, because, you know,
[33:07.400 --> 33:09.720]  it destructs it, it eats it, for instance.
[33:09.720 --> 33:11.340]  You know, humans are very good at eating
[33:11.340 --> 33:13.640]  stuff, and also other living
[33:14.060 --> 33:14.780]  beings.
[33:15.940 --> 33:17.480]  But everything that lives also has to
[33:17.480 --> 33:19.460]  die. So the more A we
[33:19.460 --> 33:21.600]  have, the more likely it is that it's also dying
[33:21.600 --> 33:23.460]  with another parameter, P4.
[33:23.760 --> 33:24.900]  And then, you know, we get
[33:25.420 --> 33:27.480]  more B, but then A gets more
[33:27.480 --> 33:29.240]  to eat, but then eats more, and then
[33:29.240 --> 33:31.240]  B goes down again. So
[33:31.240 --> 33:33.260]  there's little A that can eat up B
[33:33.260 --> 33:35.520]  and B can re-arrive, and then it's going up and down.
[33:35.520 --> 33:37.220]  So this is sort of how nowadays
[33:37.220 --> 33:39.320]  coupled ordinary differential equations,
[33:39.320 --> 33:41.340]  because there's two of them and they are depending on one
[33:41.340 --> 33:43.520]  another, we can describe the system's
[33:43.520 --> 33:45.460]  behavior and even these weird
[33:45.460 --> 33:47.580]  oscillations, like going up and down.
[33:48.320 --> 33:49.480]  Also from a mathematical
[33:49.480 --> 33:51.520]  point of view, a very interesting
[33:51.520 --> 33:53.600]  observation of this system behavior.
[33:54.360 --> 33:55.240]  But now
[33:55.240 --> 33:57.240]  I talked about A and B a lot, but
[33:57.960 --> 33:59.140]  saying B, I don't
[33:59.140 --> 34:01.220]  necessarily mean a bee. Of course, a bee is also
[34:01.380 --> 34:03.340]  a biological entity, right? The bee that gives
[34:03.340 --> 34:05.280]  you honey and stuff. But B could be
[34:05.280 --> 34:07.280]  anything, honestly. We are talking about
[34:07.280 --> 34:09.360]  the human animal, the marine animal
[34:09.360 --> 34:11.360]  like mice. It could be the number
[34:11.360 --> 34:13.260]  of cells that are changing in our
[34:13.260 --> 34:15.060]  body. It could be bacteria
[34:15.060 --> 34:17.240]  that are also in our body, also cells.
[34:17.240 --> 34:19.480]  It can be this virus, right? This is a
[34:19.480 --> 34:21.420]  picture of a virus. But it
[34:21.420 --> 34:23.380]  can also be just certain molecules that
[34:23.380 --> 34:25.180]  change over time. For instance, here an antibody
[34:25.180 --> 34:26.360]  that is produced against
[34:27.070 --> 34:28.440]  this virus.
[34:29.940 --> 34:30.540]  So
[34:31.200 --> 34:33.040]  these are all options of biological
[34:33.040 --> 34:34.960]  entities that we can describe in biological
[34:34.960 --> 34:37.300]  systems. But to describe them
[34:37.300 --> 34:39.320]  actually we have to think
[34:39.320 --> 34:41.260]  on which level are we
[34:41.260 --> 34:42.260]  talking.
[34:43.300 --> 34:45.120]  Because biology happens on
[34:45.120 --> 34:46.980]  many different scales, in space
[34:46.980 --> 34:48.920]  and in time, right? The human being is
[34:48.920 --> 34:50.460]  on the order of
[34:51.260 --> 34:51.920]  magnitude
[34:54.320 --> 34:54.800]  like
[34:54.800 --> 34:57.420]  one to two meters, let's say.
[34:58.760 --> 34:59.720]  Maybe basketball
[34:59.720 --> 35:01.460]  players are more like two meters
[35:01.460 --> 35:03.900]  and other people more like one meter.
[35:03.900 --> 35:04.800]  Never mind.
[35:05.700 --> 35:07.820]  But then we talked about
[35:08.260 --> 35:09.660]  a mouse that is certainly not a meter
[35:09.660 --> 35:11.720]  but more like centimeters.
[35:11.760 --> 35:13.500]  And then we have cells that you have
[35:13.500 --> 35:15.320]  seen that are more like micrometer
[35:15.320 --> 35:17.720]  which is a meter divided by a million.
[35:18.740 --> 35:19.060]  And then
[35:19.060 --> 35:21.240]  in these cells we have small
[35:21.740 --> 35:22.220]  subunits
[35:23.340 --> 35:24.420]  that are even
[35:25.680 --> 35:27.340]  smaller. Bacteria are also smaller
[35:27.340 --> 35:28.780]  than our body cells.
[35:29.460 --> 35:30.880]  And here we are then in the order of magnitude
[35:31.680 --> 35:32.160]  of
[35:35.780 --> 35:37.060]  a meter divided by
[35:37.340 --> 35:39.500]  a billion, okay? So ten to the minus
[35:39.500 --> 35:41.800]  nine, a nanometer.
[35:43.820 --> 35:44.340]  so... and
[35:44.340 --> 35:47.320]  all these scales, you know, changes are happening all the
[35:47.320 --> 35:48.980]  time. That's how we are being alive.
[35:48.980 --> 35:51.040]  Because molecules are turned over, cells
[35:51.040 --> 35:53.140]  are turned over, our bodies are being turned
[35:53.140 --> 35:55.080]  over. Everything happens there.
[35:55.080 --> 35:56.100]  And this not only happens
[35:57.080 --> 35:59.180]  on a very slow scale, like billions
[35:59.180 --> 36:01.140]  3.6 billion years, this earth
[36:01.140 --> 36:03.060]  does exist and life
[36:03.060 --> 36:04.340]  does exist.
[36:05.140 --> 36:07.080]  But also in a matter of hours,
[36:07.080 --> 36:09.300]  in a matter of days, in a matter of minutes, in a
[36:09.300 --> 36:11.480]  matter of seconds, you know, molecules
[36:11.480 --> 36:13.820]  change their conformation in the order
[36:13.820 --> 36:15.300]  of microseconds. Again, a
[36:15.300 --> 36:16.840]  second divided by one million.
[36:16.840 --> 36:18.740]  And this all, you know,
[36:18.740 --> 36:20.660]  this is all part of this one
[36:20.660 --> 36:22.620]  biological system that we are trying to describe
[36:22.620 --> 36:24.720]  and there's no one out there who can
[36:24.720 --> 36:26.800]  tell me that there's a way to
[36:26.800 --> 36:28.780]  handle this complexity without a computer,
[36:28.780 --> 36:30.840]  okay? So we need computer science
[36:30.840 --> 36:32.540]  to address all this, to understand
[36:32.540 --> 36:34.660]  biology at all these scales.
[36:35.080 --> 36:36.100]  And if you're curious
[36:36.640 --> 36:38.520]  for this, don't you worry
[36:38.520 --> 36:40.300]  because there's a wonderful database at
[36:40.300 --> 36:42.700]  bionumbers.org initiated
[36:42.700 --> 36:44.120]  at Harvard Medical School
[36:44.120 --> 36:45.800]  with you guys in the US
[36:45.800 --> 36:48.320]  where you can ask the question, okay, how big is a cell?
[36:48.320 --> 36:50.080]  How many molecules are in a cell?
[36:50.080 --> 36:52.160]  How does a human body cell
[36:52.160 --> 36:54.300]  is different from a bacterial body cell,
[36:54.300 --> 36:56.540]  from a fruit fly, from a
[36:56.540 --> 36:58.500]  worm? All these numbers are there
[36:58.500 --> 37:00.520]  with references. You can look up the
[37:00.520 --> 37:02.420]  original scientific literature to see
[37:02.420 --> 37:04.640]  the evidence for these numbers. And it's just
[37:04.640 --> 37:06.400]  wonderful looking at
[37:06.400 --> 37:08.280]  all this information that
[37:08.280 --> 37:10.380]  is out there. And not
[37:10.380 --> 37:12.100]  only this information is out there,
[37:12.100 --> 37:14.500]  but it's being continuously generated.
[37:14.500 --> 37:16.000]  We are living in a time now, in an
[37:16.000 --> 37:18.000]  era where biological data is
[37:18.000 --> 37:20.000]  not, for the first time, not a bottleneck
[37:20.000 --> 37:22.020]  anymore. You don't need to spend three
[37:22.020 --> 37:23.380]  years of a PhD to
[37:24.020 --> 37:25.940]  gather one data point of
[37:25.940 --> 37:28.000]  information about one protein.
[37:28.260 --> 37:29.680]  No, we can now
[37:29.680 --> 37:31.860]  sequence our
[37:31.860 --> 37:33.940]  entire genetic information, DNA
[37:33.940 --> 37:35.980]  of a patient upon
[37:35.980 --> 37:37.860]  diagnosis, after treatment, and
[37:37.860 --> 37:39.880]  really understand how our medical
[37:39.880 --> 37:40.980]  treatments affect
[37:42.460 --> 37:44.000]  our biology on so
[37:44.000 --> 37:46.060]  many levels. So we are living in
[37:46.060 --> 37:48.140]  an era of biological
[37:48.140 --> 37:50.040]  big data. But speaking to you, maybe
[37:50.040 --> 37:52.080]  of course I have to say that in fact
[37:52.080 --> 37:54.460]  there's only one true big data.
[37:54.800 --> 37:56.120]  But here I really speak
[37:56.120 --> 37:58.220]  about data
[37:58.220 --> 37:59.980]  and biology that is
[37:59.980 --> 38:02.120]  out there on all of these
[38:02.820 --> 38:03.980]  scales that I've just
[38:03.980 --> 38:05.980]  introduced to you. So
[38:05.980 --> 38:07.540]  data is not a problem anymore,
[38:07.540 --> 38:09.560]  but we need concepts how to
[38:09.560 --> 38:11.000]  work with this data.
[38:11.780 --> 38:13.700]  And in systems biology, there's two
[38:13.700 --> 38:15.540]  sort of complementary
[38:15.540 --> 38:17.540]  approaches to deal with all this
[38:17.540 --> 38:19.600]  data. So again,
[38:19.600 --> 38:21.940]  data can be on the level of organisms,
[38:21.940 --> 38:23.640]  it can be on the level of cells, or it can
[38:23.640 --> 38:25.220]  be on the level of molecules.
[38:25.740 --> 38:27.460]  But the problem is you hardly
[38:27.460 --> 38:29.740]  ever have just one of each of these.
[38:29.740 --> 38:31.460]  So usually you have many
[38:31.460 --> 38:33.520]  human beings in a clinical study, you have
[38:33.520 --> 38:36.400]  many different cells that make up our body,
[38:36.400 --> 38:38.280]  like 30 trillion of them
[38:38.280 --> 38:39.960]  to be precise, or many
[38:39.960 --> 38:42.120]  molecules that do something, right?
[38:42.120 --> 38:43.780]  Many virus particles
[38:43.780 --> 38:45.860]  or anything of that like.
[38:45.980 --> 38:48.320]  So we have two different ways to approach them.
[38:48.320 --> 38:49.980]  Either you do it sort of
[38:49.980 --> 38:52.100]  bottom up, where you say, okay,
[38:52.100 --> 38:53.560]  let's just look at
[38:54.270 --> 38:55.860]  some of these different molecules
[38:55.860 --> 38:58.080]  where we know actually what they are doing,
[38:58.080 --> 39:00.160]  how they interact with one another.
[39:00.820 --> 39:02.120]  And then we try to
[39:02.120 --> 39:04.240]  sort of describe this
[39:04.240 --> 39:06.060]  in a dynamic way,
[39:06.060 --> 39:08.120]  so how they change over time. So first one
[39:08.120 --> 39:10.160]  molecule interacts with another one, and then
[39:10.160 --> 39:12.880]  these two interact with the third one.
[39:13.100 --> 39:14.320]  And then we understand the small
[39:14.320 --> 39:16.520]  part, like we start from the ground,
[39:16.520 --> 39:18.180]  we understand the small part of it,
[39:18.180 --> 39:20.100]  the small module, and then we sort of
[39:20.100 --> 39:21.440]  subsequently try to
[39:21.980 --> 39:23.700]  connect these modules.
[39:24.360 --> 39:26.100]  And as I said, there,
[39:26.100 --> 39:28.280]  based on a lot of knowledge, what is going on,
[39:28.280 --> 39:29.460]  we have some mechanistic
[39:30.880 --> 39:32.140]  understanding how these gear
[39:32.140 --> 39:34.060]  wheels really fit into one another,
[39:34.060 --> 39:36.020]  and then we can see how this system is changing
[39:36.020 --> 39:36.960]  over time.
[39:37.940 --> 39:39.820]  But then there's another way to
[39:39.820 --> 39:42.120]  look at a biological system,
[39:42.120 --> 39:44.620]  let's say a patient, a COVID-19 patient.
[39:44.680 --> 39:46.220]  Just look at them from the
[39:46.220 --> 39:48.040]  top, okay? You don't really know what's going on
[39:48.040 --> 39:49.980]  in this patient, okay? It's a black box,
[39:49.980 --> 39:52.100]  this patient. But now you give a certain
[39:52.100 --> 39:54.000]  drug to this patient, this is
[39:54.000 --> 39:56.040]  your input into the system, into your black
[39:56.040 --> 39:58.220]  box, and you see how the patient reacts
[39:58.220 --> 40:00.160]  to it. And this is your output, your reaction.
[40:00.160 --> 40:02.120]  And then you do this not only with one, but with
[40:02.120 --> 40:03.780]  the entire cohort of
[40:03.780 --> 40:06.060]  COVID-19 patients, and you try
[40:06.060 --> 40:08.220]  to see patterns, okay? Whenever we give
[40:08.220 --> 40:10.160]  this to a certain patient, where maybe we have
[40:10.160 --> 40:12.220]  all these health records,
[40:12.780 --> 40:14.200]  there's always a good outcome.
[40:14.200 --> 40:16.280]  Or maybe not a good outcome. And then you can start
[40:16.280 --> 40:18.380]  to try to understand this input-
[40:18.380 --> 40:20.120]  output relationship. Needless to say,
[40:20.120 --> 40:22.220]  for this you need a lot of data, many patients,
[40:22.220 --> 40:23.440]  many drugs that you
[40:24.280 --> 40:25.480]  potentially test.
[40:25.800 --> 40:28.280]  But then you can learn associations.
[40:28.280 --> 40:30.340]  Okay, smoking
[40:30.340 --> 40:32.260]  history is always connected
[40:32.260 --> 40:34.400]  to a bad outcome, or as we said, a high BMI
[40:34.400 --> 40:36.140]  is always connected to a bad outcome.
[40:36.220 --> 40:38.220]  And these are all these machine learning, artificial
[40:38.220 --> 40:39.740]  intelligence methods.
[40:40.920 --> 40:42.380]  Okay, so these
[40:42.380 --> 40:44.640]  are sort of the two complementaries,
[40:44.640 --> 40:46.240]  not schools of thoughts, but approaches
[40:46.240 --> 40:48.340]  that are not mutually exclusive, right?
[40:48.340 --> 40:50.200]  Because you can, while looking at these
[40:50.200 --> 40:52.480]  patients as a black box, you can still start
[40:52.480 --> 40:54.300]  with some, for instance, cells of
[40:54.300 --> 40:56.360]  the immune system that are attacked by the virus, you can also
[40:56.360 --> 40:58.320]  start with some bottom-up modeling there, and then you
[40:58.320 --> 41:00.340]  can again put it all together.
[41:01.440 --> 41:02.520]  But to give you
[41:03.280 --> 41:04.440]  a precise example,
[41:04.440 --> 41:06.580]  I want to use the bottom-up
[41:06.580 --> 41:08.560]  approach now as an example, okay?
[41:08.560 --> 41:10.300]  So again, where we look at how
[41:10.300 --> 41:13.000]  biological entities change over time.
[41:13.240 --> 41:14.460]  And actually these models,
[41:14.460 --> 41:16.380]  particularly also with these going ups and
[41:16.380 --> 41:18.180]  down, these oscillations,
[41:18.180 --> 41:20.360]  they were introduced a long time
[41:20.360 --> 41:21.780]  ago for
[41:23.300 --> 41:23.820]  for
[41:25.420 --> 41:26.020]  foxes
[41:26.020 --> 41:27.900]  and rabbits,
[41:28.580 --> 41:30.120]  but biology is not just
[41:30.120 --> 41:32.320]  that. And in fact, I wrote a
[41:34.140 --> 41:34.740]  book
[41:36.240 --> 41:36.840]  on
[41:36.840 --> 41:38.260]  rabbit-free
[41:38.260 --> 41:40.200]  biology. However, we can still
[41:40.200 --> 41:42.440]  use all this mathematics that was established
[41:42.440 --> 41:44.240]  to now bring
[41:44.240 --> 41:46.160]  this research back to the future,
[41:46.160 --> 41:48.120]  using this mathematics to understand something
[41:48.120 --> 41:50.440]  about molecular biology.
[41:50.440 --> 41:52.020]  And the last example
[41:52.020 --> 41:53.740]  I want to give you,
[41:53.740 --> 41:55.960]  before I hand over back to Noah,
[41:55.960 --> 41:59.200]  is the example of a
[41:59.200 --> 42:00.420]  molecule that controls
[42:00.420 --> 42:02.160]  the survival of
[42:02.760 --> 42:05.240]  our red blood cells.
[42:05.240 --> 42:06.320]  It has this complicated
[42:06.320 --> 42:07.980]  name, signal transducer and activator
[42:07.980 --> 42:12.360]  of transcription, STAT5.
[42:12.360 --> 42:14.180]  The only thing you need to know about this molecule
[42:14.180 --> 42:16.380]  is it can bind
[42:16.380 --> 42:17.880]  DNA, okay? It can bind our
[42:17.880 --> 42:19.260]  genetic information and then
[42:20.140 --> 42:21.340]  exert certain
[42:21.340 --> 42:23.400]  tasks there. And if
[42:23.400 --> 42:25.360]  we have it, our red blood cells are
[42:25.360 --> 42:27.360]  surviving, and if this
[42:27.360 --> 42:28.700]  molecule is not
[42:29.140 --> 42:31.660]  at our genetic information doing its job,
[42:31.660 --> 42:33.400]  then red blood cells
[42:33.400 --> 42:35.300]  are dying. And so do we
[42:35.300 --> 42:37.520]  at some point. So this molecule
[42:37.520 --> 42:39.300]  basically shows us
[42:39.440 --> 42:41.280]  a life-death decision,
[42:41.280 --> 42:43.420]  which starts with a simple,
[42:43.420 --> 42:45.320]  very simple question. Does 1
[42:45.320 --> 42:47.280]  plus 1 always give 2? Because
[42:47.280 --> 42:49.160]  if we have two of these molecules, they are supposed
[42:49.160 --> 42:51.280]  to form this complex of two molecules.
[42:51.940 --> 42:53.180]  And this sounds simple, but
[42:53.180 --> 42:55.220]  actually biology is more complex than that, because
[42:55.220 --> 42:56.700]  we not only have one
[42:57.120 --> 42:59.200]  family member of this molecule, we have two
[42:59.200 --> 43:01.520]  of them, the blue one and the red one.
[43:01.900 --> 43:03.640]  And also this can be modified.
[43:03.640 --> 43:05.160]  It can be chemically modified to be
[43:05.720 --> 43:07.260]  switched. So this
[43:07.260 --> 43:09.280]  small p here is a chemical group
[43:09.280 --> 43:11.040]  that can be added to this molecule,
[43:11.040 --> 43:13.200]  and then it's switched from the off state
[43:13.820 --> 43:15.440]  by this addition of this chemical
[43:15.440 --> 43:16.700]  to the on state.
[43:17.640 --> 43:19.780]  And so we basically have four different
[43:20.580 --> 43:21.460]  molecules that
[43:21.460 --> 43:23.400]  potentially can do this job of finding
[43:23.400 --> 43:25.360]  genetic information and making sure
[43:25.360 --> 43:26.860]  our blood cells are surviving,
[43:27.320 --> 43:29.520]  which gives us a matrix of potential
[43:29.520 --> 43:31.960]  biological entities here.
[43:32.300 --> 43:33.640]  Not this kind of matrix,
[43:33.640 --> 43:35.440]  but more like this kind of matrix.
[43:35.820 --> 43:37.380]  So from all these molecules, we can
[43:37.380 --> 43:39.100]  form these potential complexes. And now
[43:39.100 --> 43:41.320]  the problem is they are very small, so we cannot just
[43:41.320 --> 43:43.460]  look at them to see which of them do actually exist
[43:43.460 --> 43:45.320]  and what do they do.
[43:46.240 --> 43:47.380]  But we do have
[43:47.380 --> 43:48.380]  experimental data
[43:49.140 --> 43:51.420]  where we see how these complexes change over
[43:51.420 --> 43:53.680]  time. So what we see here
[43:53.680 --> 43:55.560]  on the top is basically we see
[43:55.560 --> 43:57.460]  how many of the switched on molecules
[43:57.460 --> 43:59.420]  we have as compared to all
[43:59.420 --> 44:01.480]  possible molecules. And then
[44:01.480 --> 44:03.520]  we also see how the number of the blue
[44:03.520 --> 44:05.400]  molecules compared to all molecules change
[44:05.400 --> 44:07.560]  over time. And then here we basically
[44:07.560 --> 44:09.400]  see how these two
[44:09.400 --> 44:11.440]  ratios change with
[44:11.440 --> 44:12.420]  respect to one another
[44:12.420 --> 44:15.140]  in these cells.
[44:15.380 --> 44:17.600]  We can measure this, and then with a mathematical
[44:17.600 --> 44:19.100]  model, we can basically describe
[44:19.840 --> 44:21.200]  this data.
[44:21.280 --> 44:23.160]  As you can see, again, we have our
[44:23.160 --> 44:25.260]  simulation and also the confidence
[44:25.260 --> 44:27.420]  interval here. And now with
[44:27.420 --> 44:29.460]  the simulation, we can basically look back and see
[44:30.380 --> 44:31.640]  which kind of molecules
[44:31.640 --> 44:33.300]  from this matrix do actually
[44:33.300 --> 44:35.440]  exist. I mean, which one of those
[44:35.440 --> 44:37.480]  we need to describe this data.
[44:37.680 --> 44:39.400]  And the answer is surprisingly few
[44:39.400 --> 44:40.660]  of them we only need.
[44:41.240 --> 44:43.420]  And this is something we cannot just look at.
[44:43.420 --> 44:45.520]  We really needed the
[44:45.520 --> 44:47.360]  computer to find out what's the probability
[44:47.360 --> 44:49.020]  that these molecules exist.
[44:49.300 --> 44:51.040]  And this we also published. It's not
[44:51.040 --> 44:53.200]  an A work, but again, A and B because
[44:53.200 --> 44:55.360]  these are two molecules. And with this, maybe
[44:55.360 --> 44:57.400]  we better understand now how our
[44:57.400 --> 44:59.000]  red blood cells are surviving and
[44:59.000 --> 45:00.300]  be happy.
[45:00.980 --> 45:03.020]  To wrap this all up,
[45:03.020 --> 45:05.280]  what I've shown to you is basically there are
[45:05.280 --> 45:07.160]  standing parallels between computer viruses
[45:07.160 --> 45:09.000]  and biological viruses.
[45:09.060 --> 45:11.400]  We can apply computational algorithms
[45:11.400 --> 45:13.380]  to address challenges that we are facing
[45:13.380 --> 45:15.260]  in the COVID-19 pandemic.
[45:15.600 --> 45:17.200]  And systems biology helps us to answer
[45:17.200 --> 45:19.160]  complex biological questions with quantitative
[45:19.160 --> 45:21.260]  data, for instance, how
[45:21.260 --> 45:23.200]  molecular switches decide between life
[45:23.200 --> 45:25.320]  and death in our blood.
[45:25.480 --> 45:26.840]  And Noah will give you
[45:26.840 --> 45:28.260]  some more teasers
[45:29.020 --> 45:31.320]  how this all relates.
[45:31.320 --> 45:33.120]  I'm just saying, if you
[45:33.120 --> 45:35.280]  feel intrigued by this,
[45:35.280 --> 45:37.120]  don't hesitate to reach out to us.
[45:37.120 --> 45:39.220]  We will be happy to also
[45:39.220 --> 45:40.900]  share more information.
[45:40.900 --> 45:43.560]  Having said that, I hand over back to Noah.
[45:49.430 --> 45:51.450]  Hi, and welcome back.
[45:51.450 --> 45:52.290]  For anyone that
[45:54.350 --> 45:55.170]  joined us
[45:55.170 --> 45:57.230]  after my introduction, I will just
[45:57.230 --> 45:58.290]  explain again that
[45:59.810 --> 46:01.050]  my name is Noah.
[46:01.050 --> 46:02.890]  I'm a Junior Master Scientist at the Weizmann
[46:02.890 --> 46:04.650]  Institute of Science
[46:04.650 --> 46:06.270]  here in Israel.
[46:06.270 --> 46:07.990]  And my background
[46:07.990 --> 46:09.190]  is from...
[46:10.370 --> 46:11.830]  My background is
[46:11.830 --> 46:13.830]  as most of you guys,
[46:13.830 --> 46:16.050]  I was a reverse engineering and
[46:16.050 --> 46:18.150]  cybersecurity expert at
[46:18.150 --> 46:20.130]  Checkpoint for many years. And when I
[46:20.130 --> 46:22.050]  joined Weizmann, I had to close
[46:23.350 --> 46:24.030]  a huge
[46:24.030 --> 46:26.010]  gap in my knowledge
[46:26.010 --> 46:27.230]  in biology.
[46:27.970 --> 46:29.410]  And I was...
[46:29.410 --> 46:31.870]  What was stunning for me was the parallels
[46:31.870 --> 46:33.950]  that I found so often in my
[46:33.950 --> 46:35.490]  in the material that I
[46:35.490 --> 46:37.530]  studied in the courses and also for the lab
[46:38.570 --> 46:39.150]  with
[46:39.720 --> 46:42.230]  the knowledge that I have from the past.
[46:42.290 --> 46:44.210]  So, I will just give you
[46:44.670 --> 46:45.710]  a few examples of
[46:45.710 --> 46:48.050]  the things we have in biology
[46:48.050 --> 46:50.090]  that will look very
[46:50.090 --> 46:52.310]  familiar for you.
[46:52.490 --> 46:53.290]  So,
[46:53.840 --> 46:55.990]  without any further
[46:57.610 --> 46:58.570]  introduction,
[46:58.570 --> 46:59.610]  let's start with
[46:59.610 --> 47:01.870]  cancer cells. Okay, so cancer
[47:01.870 --> 47:03.790]  cells, unlike
[47:03.790 --> 47:06.050]  most of the people think about them,
[47:06.510 --> 47:07.770]  they are not just cancer
[47:07.770 --> 47:09.890]  that accumulate in our body. This is
[47:09.890 --> 47:10.850]  more like
[47:12.570 --> 47:13.870]  cells in our
[47:13.870 --> 47:16.070]  body that have superpowers. They evolve
[47:17.270 --> 47:17.870]  better
[47:17.870 --> 47:19.750]  than the rest of the cells in the body
[47:19.750 --> 47:21.790]  due to a mutation
[47:21.790 --> 47:23.970]  from a radiation,
[47:23.970 --> 47:25.890]  from a UV light, or from
[47:25.890 --> 47:27.890]  any other, from smoking,
[47:28.870 --> 47:29.810]  or just
[47:29.810 --> 47:32.150]  by chance in the body.
[47:32.150 --> 47:33.750]  And then, what's happening with
[47:33.750 --> 47:36.010]  these cells is that they become selfish.
[47:36.010 --> 47:37.530]  Which means that instead of
[47:37.530 --> 47:39.230]  working together
[47:39.570 --> 47:41.890]  with the rest of the cells in the body,
[47:41.890 --> 47:43.850]  they are selfish enough
[47:43.850 --> 47:45.490]  to try to control
[47:47.150 --> 47:47.890]  and to
[47:47.890 --> 47:49.870]  get all of the resources, for example,
[47:50.370 --> 47:51.570]  they are actually
[47:51.570 --> 47:53.630]  sending signals to the
[47:53.630 --> 47:55.090]  body to build
[47:55.810 --> 47:57.670]  veins in order to get more oxygen
[47:57.670 --> 48:00.790]  into the tumor environment.
[48:01.150 --> 48:01.410]  And
[48:01.410 --> 48:02.370]  one
[48:04.270 --> 48:05.790]  stunning fact that I learned
[48:05.790 --> 48:07.250]  during my courses
[48:07.250 --> 48:09.810]  is that not only that the cancer
[48:09.810 --> 48:11.790]  cells are selfish, they also
[48:11.790 --> 48:13.730]  try to evade the immune
[48:13.730 --> 48:15.670]  system, because
[48:15.670 --> 48:17.610]  one of the roles of the immune system is
[48:17.610 --> 48:19.470]  not only to
[48:19.470 --> 48:21.490]  eliminate pathogens, but also
[48:21.490 --> 48:23.730]  eliminate cells that are selfish, like
[48:23.730 --> 48:25.670]  cancer cells. Because cancer cells, like we
[48:25.670 --> 48:27.610]  have them on a regular basis in our
[48:27.610 --> 48:29.430]  body, and
[48:29.430 --> 48:31.650]  this is one of the main roles of the immune
[48:31.650 --> 48:33.610]  system, to detect
[48:33.610 --> 48:35.030]  them and
[48:35.670 --> 48:37.190]  to eliminate them.
[48:37.470 --> 48:39.550]  So, it's not
[48:39.550 --> 48:41.530]  only that our immune system
[48:41.530 --> 48:43.670]  is working as an antivirus, but
[48:43.670 --> 48:45.710]  the thing here is that the cancer cells
[48:45.710 --> 48:47.030]  are fighting back.
[48:47.410 --> 48:49.350]  So, the picture here is representing
[48:49.350 --> 48:51.730]  the interaction between cancer cells and different types
[48:51.730 --> 48:53.810]  of immune cells. Cancer cells
[48:53.810 --> 48:55.750]  are the one that shows in blue.
[48:55.750 --> 48:58.090]  Immune cells are in different colors, and
[48:58.090 --> 49:00.130]  what they are trying to do
[49:00.830 --> 49:01.810]  is to
[49:01.810 --> 49:03.890]  send... so, you can
[49:04.430 --> 49:06.650]  think about the rest of the
[49:06.650 --> 49:07.630]  colorful cells as
[49:07.630 --> 49:09.730]  different types of antivirus from
[49:09.730 --> 49:11.950]  many different companies,
[49:11.950 --> 49:13.570]  or maybe
[49:13.570 --> 49:15.770]  different models that we have in one
[49:15.770 --> 49:17.790]  antivirus. And what the cancer cell
[49:17.790 --> 49:20.250]  is trying to do is to actually reprogram
[49:20.250 --> 49:21.810]  or to send
[49:23.090 --> 49:23.610]  signals
[49:25.150 --> 49:26.710]  to different immune cells
[49:26.710 --> 49:28.690]  in order to evade
[49:28.690 --> 49:30.850]  them. And this is something
[49:30.850 --> 49:32.810]  that was, you know, when I heard about it,
[49:32.810 --> 49:34.910]  it was really familiar for me. Like,
[49:34.910 --> 49:36.650]  there were many models that tried
[49:36.650 --> 49:39.590]  to evade my VM, to evade
[49:39.590 --> 49:40.270]  my
[49:41.390 --> 49:42.610]  AIDA, my
[49:42.610 --> 49:44.770]  OLADPG, or the antivirus
[49:44.770 --> 49:46.270]  in general. And
[49:46.270 --> 49:48.870]  so, for me, it sounds pretty familiar.
[49:48.870 --> 49:49.850]  So, apparently,
[49:49.850 --> 49:51.810]  it wasn't our
[49:51.810 --> 49:53.230]  idea to evade
[49:53.850 --> 49:55.770]  people, to evade
[49:56.950 --> 49:57.950]  entities that
[49:57.950 --> 49:59.830]  are trying to detect us.
[49:59.830 --> 50:01.770]  And I can promise you that cancer
[50:01.770 --> 50:02.670]  cells are doing it in
[50:04.150 --> 50:06.190]  a much more efficient way.
[50:06.270 --> 50:08.050]  For example, one
[50:08.050 --> 50:09.650]  very stunning
[50:10.430 --> 50:11.870]  example that I have
[50:12.590 --> 50:13.650]  is that
[50:13.650 --> 50:15.650]  every cell is supposed to
[50:15.650 --> 50:16.970]  present
[50:18.730 --> 50:19.290]  a molecule
[50:19.290 --> 50:20.830]  on its surface, which
[50:20.830 --> 50:23.590]  the name of the molecule is an MHC.
[50:23.610 --> 50:25.270]  And this molecule
[50:25.270 --> 50:26.250]  is actually like
[50:27.370 --> 50:28.490]  the info
[50:28.490 --> 50:30.950]  that an executable
[50:30.950 --> 50:32.570]  can have, which
[50:32.570 --> 50:34.790]  describes what
[50:34.790 --> 50:37.250]  these cells are producing,
[50:37.250 --> 50:39.070]  what are the proteins that are running inside
[50:39.070 --> 50:41.010]  of it. And cancer cells,
[50:41.010 --> 50:43.130]  what they actually do is they just
[50:43.130 --> 50:44.770]  don't show, they just don't present
[50:44.770 --> 50:46.870]  these MHC molecules,
[50:47.290 --> 50:48.570]  so it's really hard
[50:48.570 --> 50:50.890]  for the immune cells
[50:50.890 --> 50:52.890]  to understand what actually is going on
[50:52.890 --> 50:54.150]  inside of them.
[50:55.530 --> 50:55.770]  And
[50:58.370 --> 50:58.970]  here
[50:58.970 --> 51:00.790]  you can see something that
[51:00.790 --> 51:02.710]  we call a hot tumor.
[51:02.710 --> 51:04.790]  If a tumor is immunologically
[51:04.790 --> 51:06.690]  hot, it means that the immune system
[51:06.690 --> 51:09.090]  is operating here. However,
[51:09.090 --> 51:10.890]  when cancer progresses, it learns
[51:10.890 --> 51:12.830]  how to deactivate immune cells
[51:12.830 --> 51:14.690]  and make tumors
[51:14.690 --> 51:16.970]  immunologically cold. Like, look at this
[51:16.970 --> 51:18.890]  tumor, and we can
[51:18.890 --> 51:20.490]  see here many different
[51:20.490 --> 51:20.790]  of
[51:22.710 --> 51:24.930]  immune cells that are trying to detect
[51:24.930 --> 51:26.470]  and trying to
[51:26.470 --> 51:28.830]  get attached to it and to understand what's going
[51:28.830 --> 51:31.050]  on inside. But after a
[51:31.050 --> 51:33.310]  while, what we can see in tumors
[51:33.310 --> 51:35.110]  that are actually
[51:35.110 --> 51:37.110]  successful with evading the immune system
[51:37.110 --> 51:39.690]  is that they become immunologically
[51:39.690 --> 51:40.810]  cold tumors. You can see
[51:40.810 --> 51:41.970]  that most of the
[51:42.790 --> 51:44.570]  immune cells just gave up
[51:44.570 --> 51:45.230]  and
[51:46.630 --> 51:48.510]  they said, okay, whatever,
[51:48.510 --> 51:50.850]  I have other stuff to do, and they just left
[51:50.850 --> 51:52.350]  the tumor there alone.
[51:52.690 --> 51:53.410]  This is why
[51:55.690 --> 51:56.510]  you get
[51:57.030 --> 51:59.150]  a tumor in your body.
[51:59.610 --> 52:01.050]  And here is just
[52:01.050 --> 52:02.770]  what you can see here
[52:02.770 --> 52:03.410]  is just
[52:05.550 --> 52:06.490]  a network of
[52:07.090 --> 52:09.210]  different... this is just
[52:09.370 --> 52:10.190]  a brief
[52:11.070 --> 52:13.170]  knowledge that we have so far about
[52:15.770 --> 52:16.630]  all the
[52:17.210 --> 52:19.330]  different ways that cancer
[52:20.150 --> 52:21.390]  has in order
[52:21.390 --> 52:23.490]  to try to evade the immune system.
[52:23.590 --> 52:25.210]  Generally, cells can communicate
[52:25.210 --> 52:26.530]  with each other using diverse
[52:27.210 --> 52:29.170]  signal transaction networks
[52:29.170 --> 52:31.690]  that lead to different outcomes.
[52:31.690 --> 52:33.070]  And what cancer cells do, they
[52:33.070 --> 52:35.210]  upregulate the pathways which
[52:35.210 --> 52:36.530]  inhibit normal
[52:37.350 --> 52:38.850]  function of immune cells.
[52:38.850 --> 52:40.710]  They don't even
[52:40.710 --> 52:41.610]  just
[52:42.930 --> 52:43.870]  tell the
[52:44.970 --> 52:46.730]  specific immune cell, you know, just leave
[52:46.730 --> 52:48.890]  me alone. They also ask him to spread
[52:48.890 --> 52:50.870]  the rumor and tell
[52:50.870 --> 52:52.650]  the rest of the immune cells, you know,
[52:52.650 --> 52:54.730]  this guy, it's okay,
[52:54.730 --> 52:56.450]  you can just leave him alone.
[52:56.910 --> 52:58.510]  And this is something that was
[52:58.510 --> 53:00.530]  really fascinating for me to learn.
[53:01.650 --> 53:02.590]  And here
[53:02.590 --> 53:03.570]  is an example
[53:04.230 --> 53:05.490]  where cancers are
[53:05.490 --> 53:07.830]  overexposing PD-L1,
[53:07.910 --> 53:10.170]  a molecule that suppresses T-cell.
[53:10.170 --> 53:11.570]  Important cell type involved
[53:11.570 --> 53:13.290]  in cancer cell elimination.
[53:13.570 --> 53:15.630]  On the right, you can also see
[53:15.850 --> 53:17.550]  a reprogrammed immune cell called
[53:17.550 --> 53:19.590]  tumor-associated macrophage,
[53:19.590 --> 53:20.890]  which helps to spread
[53:20.890 --> 53:23.590]  the inactivating signal instead of
[53:23.590 --> 53:25.430]  fighting cancer. Okay, so
[53:25.430 --> 53:27.670]  you can see that it's not only telling him, listen,
[53:27.670 --> 53:30.150]  I'm okay, it also asks him,
[53:30.150 --> 53:31.730]  it also asks these cells,
[53:31.730 --> 53:33.790]  you can also tell the rest of the guys
[53:33.790 --> 53:36.110]  that I'm fine and to leave me alone.
[53:36.190 --> 53:37.830]  And this was amazing.
[53:37.830 --> 53:39.850]  I don't know too many models that are
[53:39.850 --> 53:41.670]  actually so
[53:42.410 --> 53:43.650]  powerful and sophisticated
[53:43.650 --> 53:46.070]  and it was amazing for me to learn.
[53:46.870 --> 53:47.950]  And my next
[53:47.950 --> 53:49.730]  example, which is actually one of my
[53:50.770 --> 53:52.490]  favorites, is bacteriophage.
[53:52.830 --> 53:53.830]  Okay, so
[53:53.830 --> 53:55.730]  what is a bacteriophage? Here you
[53:55.730 --> 53:57.250]  can see,
[53:57.250 --> 53:59.630]  let's take a look
[53:59.630 --> 54:01.090]  at this very beautiful
[54:02.670 --> 54:03.150]  gif
[54:04.490 --> 54:04.970]  of
[54:06.390 --> 54:07.530]  like, okay, so
[54:07.530 --> 54:09.610]  I'll just explain. First of all, bacteriophage
[54:09.610 --> 54:12.090]  is like my favorite creature so far.
[54:12.570 --> 54:13.730]  And what it's actually
[54:13.730 --> 54:15.490]  doing, there is a creature
[54:15.490 --> 54:17.570]  that we all know, it's known for bacteria.
[54:17.710 --> 54:19.530]  But there is also another tiny
[54:19.530 --> 54:20.830]  picture and
[54:22.290 --> 54:23.650]  what bacteriophage is actually
[54:23.650 --> 54:24.990]  trying to do is to
[54:25.910 --> 54:27.310]  inject its DNA
[54:27.930 --> 54:29.270]  into the bacteria
[54:29.270 --> 54:31.450]  and then the bacteria, instead of
[54:32.490 --> 54:33.490]  running their
[54:33.490 --> 54:35.550]  own genome, it will run
[54:35.550 --> 54:37.630]  the bacteriophage genome.
[54:37.630 --> 54:39.570]  Does it sound familiar to anyone
[54:39.570 --> 54:41.570]  else? Apparently also with this we didn't
[54:41.570 --> 54:43.490]  come up. And this
[54:43.490 --> 54:45.290]  was amazing for me to understand.
[54:45.290 --> 54:47.510]  You can actually see it. This is actually
[54:47.730 --> 54:49.190]  a bacteria,
[54:49.190 --> 54:51.650]  and here you can see a bacteriophage
[54:51.650 --> 54:53.230]  that are attaching
[54:53.230 --> 54:55.650]  into it and injecting
[54:55.650 --> 54:57.870]  their genome into the
[54:57.870 --> 54:59.630]  bacteria. And here you
[54:59.630 --> 55:01.630]  can see this is a very sad story because the
[55:01.630 --> 55:03.650]  bacteria, sometimes they
[55:03.650 --> 55:04.510]  understand that they are
[55:06.430 --> 55:07.310]  doomed.
[55:07.450 --> 55:08.930]  But this is just a gif.
[55:09.270 --> 55:11.890]  So I will just explain a bit about
[55:12.630 --> 55:13.750]  the progress of it.
[55:13.750 --> 55:15.630]  So what you see here is that there is
[55:15.630 --> 55:17.590]  attachment of the bacteriophage into
[55:17.590 --> 55:19.030]  the bacteria. Bacteria is
[55:20.270 --> 55:21.430]  the bacteria that
[55:21.430 --> 55:24.070]  you know, that we all have in our body.
[55:24.070 --> 55:25.330]  And this is the bacteriophage
[55:25.890 --> 55:28.790]  which has a very, very
[55:28.790 --> 55:30.050]  nice and interesting
[55:30.050 --> 55:33.050]  picture and structure.
[55:33.070 --> 55:34.110]  So what it's doing
[55:34.110 --> 55:36.970]  is actually attaching into the bacteria.
[55:37.030 --> 55:38.070]  It's injecting
[55:40.370 --> 55:41.850]  its own genome
[55:41.850 --> 55:42.350]  into the
[55:43.230 --> 55:44.130]  bacteria's inside
[55:44.130 --> 55:46.270]  surface. And then the bacteria,
[55:46.270 --> 55:48.050]  instead of running, like you see, this is the
[55:48.050 --> 55:50.150]  genome of the bacteria.
[55:50.150 --> 55:52.410]  Instead of running its own code,
[55:52.410 --> 55:54.530]  it's running the bacteriophage code.
[55:54.690 --> 55:56.470]  And then the whole purpose of it
[55:56.470 --> 55:56.930]  is that
[55:58.350 --> 56:00.270]  there are more and more
[56:00.270 --> 56:01.590]  bacteriophages that are
[56:02.650 --> 56:05.030]  being formed inside of the bacteria.
[56:05.190 --> 56:05.650]  And then
[56:06.210 --> 56:08.590]  the poor bacteria, you know,
[56:08.590 --> 56:10.210]  are just full of
[56:10.210 --> 56:12.870]  bacteriophages until they release each other.
[56:12.870 --> 56:14.590]  Until they get, like,
[56:14.590 --> 56:16.310]  it's too much for it.
[56:16.310 --> 56:17.730]  And they just go out
[56:17.730 --> 56:19.590]  and infect even more and more
[56:20.270 --> 56:21.610]  bacteria that are inside
[56:21.610 --> 56:22.930]  of this environment.
[56:23.970 --> 56:24.110]  And
[56:25.650 --> 56:26.670]  another thing
[56:27.210 --> 56:29.610]  that I'm pretty sure that most of you heard
[56:29.610 --> 56:31.570]  about, if you are interested in
[56:31.570 --> 56:33.470]  biology at all, is
[56:33.470 --> 56:35.710]  something that is called CRISPR.
[56:35.710 --> 56:37.670]  So we all know that CRISPR is being
[56:37.670 --> 56:39.190]  used these days in order to
[56:39.750 --> 56:41.730]  kind of engineering the babies
[56:41.730 --> 56:43.510]  of the future and stuff like this. But what was
[56:43.510 --> 56:45.410]  fascinating for me to learn is that
[56:45.410 --> 56:47.630]  actually CRISPR, the whole
[56:47.630 --> 56:49.290]  purpose of it, is
[56:49.290 --> 56:51.510]  this is a defense
[56:51.510 --> 56:53.570]  mechanism of the bacteria
[56:53.570 --> 56:55.410]  in order to detect bacteria,
[56:55.410 --> 56:58.050]  in order to detect phages.
[56:58.050 --> 56:59.650]  So what you see here is
[56:59.650 --> 57:00.990]  that this
[57:01.630 --> 57:03.630]  pink evil thing is
[57:03.630 --> 57:05.610]  the bacteria. And what
[57:05.610 --> 57:07.610]  it's trying to detect, and
[57:07.610 --> 57:09.650]  this is something that we actually
[57:09.650 --> 57:11.530]  stole from them,
[57:11.530 --> 57:13.510]  is what it's trying to detect is
[57:13.510 --> 57:15.250]  the sequence that he also
[57:15.250 --> 57:16.090]  that he
[57:18.290 --> 57:19.370]  saw before
[57:19.370 --> 57:21.270]  of phages that try to
[57:21.270 --> 57:23.130]  attack it. And the way
[57:23.130 --> 57:25.150]  that it's doing this
[57:25.150 --> 57:27.150]  is just, you know, he has a sequence
[57:27.150 --> 57:28.930]  that he remembers
[57:28.930 --> 57:31.390]  and then he just tries to
[57:31.390 --> 57:33.570]  do pairing, like to compare
[57:33.570 --> 57:35.710]  the new
[57:35.710 --> 57:37.030]  DNA that he sees
[57:37.030 --> 57:39.150]  inside of the cell
[57:39.150 --> 57:39.830]  and
[57:41.050 --> 57:42.870]  the sequence that he already knows that
[57:42.870 --> 57:44.810]  this belongs to the phages.
[57:46.350 --> 57:54.330]  So, of course, I think
[57:54.330 --> 57:55.570]  it's very
[57:56.650 --> 57:57.850]  obvious that
[57:59.030 --> 58:00.390]  this Cas9
[58:00.390 --> 58:02.450]  system is very similar to what we know as
[58:02.450 --> 58:04.290]  MD5, or the hashes of the
[58:04.290 --> 58:05.850]  models that we have inside
[58:07.110 --> 58:08.470]  many different
[58:08.470 --> 58:09.790]  antiviruses.
[58:11.070 --> 58:11.930]  And another
[58:11.930 --> 58:13.910]  very cool thing is that it's not
[58:13.910 --> 58:16.230]  only that the bacteria is detecting
[58:16.230 --> 58:17.870]  phages, like
[58:18.610 --> 58:19.910]  using the known
[58:19.910 --> 58:21.850]  sequence, is that the phages
[58:21.850 --> 58:24.170]  are actually invading this defensive
[58:24.170 --> 58:26.370]  system by changing their genome.
[58:26.370 --> 58:28.070]  The same way that malwares are changing
[58:28.070 --> 58:30.250]  their MD5, you know, it's
[58:30.250 --> 58:32.610]  enough to change just one bit.
[58:32.690 --> 58:33.990]  Phages are doing the
[58:33.990 --> 58:35.010]  same in order to
[58:36.490 --> 58:38.290]  like, you know, via evolution
[58:38.290 --> 58:39.990]  in order to being able
[58:39.990 --> 58:42.050]  to inject
[58:42.050 --> 58:44.190]  their code into new bacterias.
[58:44.950 --> 58:45.470]  And
[58:45.990 --> 58:47.890]  so this mechanism that we
[58:47.890 --> 58:49.730]  actually stole from bacterias, we will
[58:50.810 --> 58:51.930]  we are trying
[58:51.930 --> 58:53.730]  these days to use it for genetic
[58:53.730 --> 58:55.510]  engineering, but
[58:55.510 --> 58:56.670]  just so you know,
[58:56.670 --> 58:59.870]  it wasn't our idea, we stole it
[58:59.870 --> 59:01.910]  from a very actually
[59:01.910 --> 59:03.950]  smart creature, apparently.
[59:04.510 --> 59:05.270]  And I will just
[59:06.150 --> 59:07.850]  since our time is up,
[59:07.850 --> 59:08.850]  I will just
[59:10.010 --> 59:11.370]  say briefly what
[59:11.370 --> 59:13.850]  the role of our specific
[59:13.850 --> 59:15.590]  lab. So what you see is
[59:15.590 --> 59:18.110]  so my research
[59:18.110 --> 59:19.830]  is specifically about
[59:19.830 --> 59:22.170]  the microbiome. What is the microbiome
[59:22.170 --> 59:23.830]  is what we are trying
[59:23.830 --> 59:25.670]  to do as you all
[59:25.670 --> 59:27.870]  I don't know if you heard about it, but
[59:27.870 --> 59:29.790]  we are all full of bacterias
[59:29.790 --> 59:31.850]  inside of our gut and we
[59:31.850 --> 59:33.090]  are trying in our lab
[59:34.830 --> 59:35.870]  to understand
[59:35.870 --> 59:37.950]  and maybe in the future
[59:37.950 --> 59:39.730]  also to manipulate the different bacterias
[59:39.730 --> 59:41.630]  that are inside of our gut and all the networks
[59:41.630 --> 59:43.870]  that are inside of it. First of all
[59:43.870 --> 59:45.710]  we try to find the connection between them
[59:45.710 --> 59:46.850]  and different diseases like
[59:47.610 --> 59:49.650]  diabetes and ALS
[59:49.650 --> 59:51.310]  and obesity.
[59:52.690 --> 59:53.750]  And so this is
[59:54.010 --> 59:56.050]  a very new and interesting field
[59:56.050 --> 59:57.690]  and you should know
[59:57.690 --> 01:00:00.030]  that you all have bacterias
[01:00:00.030 --> 01:00:01.830]  all over your hands, your
[01:00:01.830 --> 01:00:04.170]  eyes and specifically inside your gut
[01:00:04.170 --> 01:00:06.130]  and it's not just that they are there
[01:00:06.130 --> 01:00:08.050]  happily not doing anything, they are
[01:00:08.050 --> 01:00:09.790]  actually sometimes very crucial
[01:00:11.710 --> 01:00:13.730]  in order to keep the
[01:00:13.730 --> 01:00:15.230]  homeostasis.
[01:00:16.310 --> 01:00:17.810]  So this was just
[01:00:18.330 --> 01:00:20.050]  a brief thing
[01:00:20.050 --> 01:00:21.670]  about the microbiome, just a brief
[01:00:21.670 --> 01:00:23.910]  introduction to the microbiome.
[01:00:23.970 --> 01:00:26.050]  Just a few things that we
[01:00:26.050 --> 01:00:28.090]  found out so far in our lab is the
[01:00:28.090 --> 01:00:29.990]  connection between the bacterias that are inside
[01:00:29.990 --> 01:00:31.990]  the gut and the very well known
[01:00:31.990 --> 01:00:33.810]  disease that are inside the ALS because
[01:00:33.810 --> 01:00:36.170]  the bacterias, they can
[01:00:36.170 --> 01:00:37.890]  send signals from the gut
[01:00:37.890 --> 01:00:39.890]  into the brain and
[01:00:39.890 --> 01:00:41.430]  there is a very
[01:00:42.130 --> 01:00:43.810]  clear connection between
[01:00:43.810 --> 01:00:46.010]  the dysbiosis that we will see in the
[01:00:46.010 --> 01:00:46.870]  gut and
[01:00:48.590 --> 01:00:50.090]  the form of this
[01:00:50.090 --> 01:00:51.990]  disease. Here you
[01:00:51.990 --> 01:00:53.270]  can see many different
[01:00:54.510 --> 01:00:56.230]  types of diseases that are linked
[01:00:57.090 --> 01:00:58.110]  to dysbiosis
[01:00:58.110 --> 01:00:59.690]  in the microbiome.
[01:00:59.690 --> 01:01:01.770]  And another very cool
[01:01:02.890 --> 01:01:03.930]  subject we are
[01:01:03.930 --> 01:01:06.150]  looking into in our lab is diabetes.
[01:01:07.190 --> 01:01:07.690]  Just
[01:01:07.690 --> 01:01:09.870]  one thing that I learned
[01:01:09.870 --> 01:01:11.650]  and it was really interesting
[01:01:11.650 --> 01:01:12.210]  and
[01:01:15.310 --> 01:01:15.810]  like
[01:01:15.810 --> 01:01:17.650]  this is one of my favorite challenges
[01:01:17.650 --> 01:01:19.850]  in this field, is that
[01:01:19.850 --> 01:01:21.830]  when you have any kind of disease, it's not like
[01:01:21.830 --> 01:01:23.870]  that, for example diabetes, if you have
[01:01:23.870 --> 01:01:25.970]  too much sugar in your blood,
[01:01:25.970 --> 01:01:27.630]  the problem with it
[01:01:27.630 --> 01:01:28.790]  wouldn't be
[01:01:29.450 --> 01:01:31.490]  just about the fact that you have too much
[01:01:31.490 --> 01:01:32.990]  sugar in your veins.
[01:01:33.390 --> 01:01:35.230]  Many different diseases,
[01:01:35.230 --> 01:01:37.370]  they are like, there is a hip
[01:01:37.370 --> 01:01:39.530]  corruption, let's say,
[01:01:39.530 --> 01:01:41.310]  in our body, but unfortunately
[01:01:41.310 --> 01:01:43.510]  unlike in new operation
[01:01:43.510 --> 01:01:45.190]  systems, in our body
[01:01:45.190 --> 01:01:47.670]  there is no such thing
[01:01:47.670 --> 01:01:49.450]  as different processes that
[01:01:49.450 --> 01:01:51.330]  have different
[01:01:51.970 --> 01:01:53.370]  classes of memory.
[01:01:53.370 --> 01:01:55.370]  So if there is hip corruption
[01:01:55.370 --> 01:01:57.090]  in one process,
[01:01:57.090 --> 01:01:59.490]  in one system
[01:01:59.490 --> 01:02:01.470]  in our body, it affects the rest
[01:02:01.470 --> 01:02:02.830]  of it as well.
[01:02:02.910 --> 01:02:05.290]  For example, in diabetes, one of the problems
[01:02:05.290 --> 01:02:07.410]  that we see is that
[01:02:09.190 --> 01:02:10.810]  we lose the
[01:02:10.810 --> 01:02:12.790]  homeostasis of
[01:02:14.570 --> 01:02:15.250]  also
[01:02:16.310 --> 01:02:16.990]  the
[01:02:16.990 --> 01:02:18.570]  CA2 inside our
[01:02:21.070 --> 01:02:22.430]  cells.
[01:02:22.470 --> 01:02:23.250]  And actually
[01:02:23.250 --> 01:02:24.830]  we can see that
[01:02:24.830 --> 01:02:27.490]  the blood is becoming acidic.
[01:02:28.110 --> 01:02:30.090]  So this is just one thing.
[01:02:31.550 --> 01:02:33.230]  so, I would just
[01:02:33.230 --> 01:02:35.350]  say briefly that I'm really proud
[01:02:35.350 --> 01:02:37.210]  really, really proud
[01:02:37.210 --> 01:02:39.250]  that I'm taking part
[01:02:39.250 --> 01:02:40.110]  of my lab.
[01:02:40.110 --> 01:02:43.230]  The very nice person that you see
[01:02:43.230 --> 01:02:45.330]  here is my supervisor. I have a professor
[01:02:45.330 --> 01:02:47.070]  who is also a supervisor, and he
[01:02:47.070 --> 01:02:48.470]  recently used me as
[01:02:49.570 --> 01:02:51.290]  a lab rat. What you see
[01:02:51.290 --> 01:02:53.450]  here is an insulin
[01:02:53.450 --> 01:02:54.650]  sensor that
[01:02:55.210 --> 01:02:57.130]  was attached to my shoulder for
[01:02:57.130 --> 01:02:58.770]  around three weeks.
[01:02:59.030 --> 01:03:01.610]  But I'm, of course, very proud.
[01:03:02.150 --> 01:03:03.210]  I wish for
[01:03:03.210 --> 01:03:05.250]  myself and for everyone always to be in the side
[01:03:05.250 --> 01:03:06.570]  to help
[01:03:08.170 --> 01:03:09.130]  patients get
[01:03:09.370 --> 01:03:11.350]  a better cure.
[01:03:11.850 --> 01:03:13.330]  And this is something that I'm really
[01:03:13.330 --> 01:03:15.210]  happy about, that what I'm
[01:03:15.210 --> 01:03:17.110]  doing here is actually helping
[01:03:17.110 --> 01:03:19.730]  people that are sick.
[01:03:19.730 --> 01:03:21.150]  But yes,
[01:03:21.510 --> 01:03:22.470]  also,
[01:03:23.430 --> 01:03:26.450]  all of us actually do stuff like this in the lab.
[01:03:26.450 --> 01:03:27.530]  Also, the reason that
[01:03:27.530 --> 01:03:29.330]  this guy is dressed like this is because
[01:03:29.330 --> 01:03:31.530]  during the corona,
[01:03:31.530 --> 01:03:33.530]  our whole lab became
[01:03:33.530 --> 01:03:34.690]  of course,
[01:03:34.690 --> 01:03:36.850]  we are not
[01:03:37.750 --> 01:03:39.570]  doing anything that is linked
[01:03:40.290 --> 01:03:40.850]  to
[01:03:42.630 --> 01:03:43.750]  corona.
[01:03:43.830 --> 01:03:45.070]  We didn't do anything
[01:03:45.470 --> 01:03:47.850]  about it like before,
[01:03:47.850 --> 01:03:48.870]  but during
[01:03:49.350 --> 01:03:51.850]  this breakdown
[01:03:51.850 --> 01:03:52.970]  of...