okay so the last century or century and
a half has been called the century of
biology Elevens essentially of genetics
because it’s really been an fantastic
and interesting adventure because as
recently as in the middle of the 19th
century we had no idea how heredity
worked
starting with Mendel’s work and that
actually entered the scientific discuss
only at the beginning of the 20th
century we started with the idea that
genes are some abstract entities that
determine certain traits of organisms
for example there is a gene that makes
plant grow tall or short there is a gene
that makes the flowers red or white
there is a gene that makes flies ice
white instead of red etc etc so abstract
entities that determine certain traits
and from that in about a century we went
to this to a very good understanding of
the very chemical and physical language
of the genes we know that genes are made
of DNA we understand the alphabet of DNA
we can read it we can even rewrite it we
can take a gene from one organism and
put it in a different organism like here
a gene format relative was put into
different animals making them
fluorescent by the way this is not just
a party trick this is a very good tool
to study where genes work in a given
organism and soon some say we might get
to modifying humans to getting
custom-designed
babies with better mine’s better bodies
better everything well not so fast it’s
not as easy
as it seems even though thanks to this
enormous development of genetics from
nothing at the beginning of the 20th
century to where we are now we know
pretty well how a gene works we
understand how ATC’s and G’s get read
gets translated into proteins that make
our cells biologic textbooks are full of
this information of course we still work
on the details but we already have a
pretty good idea about how it works
these days I could take DNA from any one
of you and treat every single letter
written in this DNA it would take a few
days and about a few hundred cures to do
it for any single individual and the
databases are full of such information I
can find genes in the DNA of any other
organism although if you look carefully
at those gene lists you will mostly see
the names of awful diseases that happens
when one of those genes goes wrong you
don’t see a gene for intelligence beauty
talent etc why is that why is that not
so simple because maybe if we can read
DNA if we understand DNA then maybe we
can realize what has for a long time
being a sci-fi vision that we can take
DNA from anyone and predict their future
their talents the possibilities that
determine the entire future life well
not so fast
let’s start with something simple
something that may be familiar from
those of you who took some genetics in
high school there is not a single gene
that determines the color of your eyes
if you believe that from your school
then it’s wrong but it’s still simple
there are about six genes that make your
eyes dark brown or bright blue when you
read and check about a dozen genes you
can make a pretty good prediction about
a person’s color of her or eyes again
not as simple as it might seem but
doable you need a dozen genes some
mathematics and you can predict whether
some of somebody’s eyes look like this
or like this but that’s still a simple
example let’s take something more
challenging how tall you are there is
not a single gene that makes you tall or
short to understand why some people grow
real tall and some not so much you need
to look into at least 200 different
genes and I can tell you that at the
moment we don’t have enough mathematical
theory of how those genes interact to
make this prediction I cannot take your
DNA and predict all you should be about
6 feet tall no it is not possible it can
be done for 4 traits that depend on a
dozen genes but not on those that depend
on 200 genes and that’s all not yet that
complex because if we look at some other
traits
is there a gene that makes you behave in
a certain way for example is there a
gene that makes you more likely to enjoy
dangerous behaviors like jumping from a
perfectly good airplane there were
headlines in the newspapers that there
is a gene that makes you a risk-taker
no it’s not true complex behaviors
depend on many genes there is not a
single risk-taking gene there is not a
single intelligence genes there is no
smart gene at least a thousand or two
thousand genes would be necessary to
explain the hereditary basics of of
intelligence there is no not a single
gene
that makes you susceptible to common
diseases like heart diseases like most
cancers like autoimmune diseases etc
there are some diseases that depend on
single genes but they are very very rare
the common ones depends on thousands of
different genes so the common theme is
that you cannot explain the traits of
organisms by looking at single genes you
need to look at the interactions between
many genes and that’s a very common
theme in the entire field of biology
that complexity is built from
interaction of simple elements like
those Lego bricks it’s not hard to
understand a single element those bricks
children understands them but the
complexity lies in the interactions
between those elements human genome has
about 20,000 genes so 20,000 bricks and
the complexity of this is not the
complexity of the single elements it’s
their combinations it’s their
interactions so in need if you want to
understand the whole you need to
understand not just the elements but how
they interact and how can we approach
those complex interactions in genetics
one of the ways to look at it is to
realize that interacting elements form
networks and the world is made of
networks here you can see airline
connection networks anybody who’s
planned a longer journey has experienced
this you can see that these are not just
random connections there are hubs that
are connected to each other and
peripheries that are connected to hubs
and this can be described in a certain
formal manner your brain is also a
network it’s composed of a hundred
billion elements called neurons and we
start to map all those interactions
there is a project so
human connectome project to understand
all those connections and the elements
are quite simple they are called neurons
or nerve cells and we understand them
already pretty well we can model how a
single nerve cell behaves but we take a
hundred billion of them and we have your
mind unpredictable and complex the
complexity is built on interaction and
knowing the elements is not enough to
know the entire system another example
social networks people interact with
each other and these days everybody does
it online that’s an example of a Twitter
followers Network that’s Facebook
connections and if you look at those
this picture it may resemble what you’ve
seen on this airline connection map
there are also cups and peripheries
everybody has this friend who knows
thousands of people and when you see who
is friends with those central hubs you
can find out that few steps are enough
to connect you to about anyone there is
a theory that says that six handshakes
six steps are enough to connect any two
people in the entire world if the world
is made of networks we need some means
some way to understand to describe
networks and we can do it in biology as
well so let’s make a Facebook of Dean’s
let’s make a social network of genes
that’s a picture from a recent paper
showing interactions between genes in a
simple yeast cell as you can see the
scientists map mapped all the
interactions it had on it has only about
6,000 genes so less than a human cell
and you can see that there are genes
that are hubs that have many connections
and there are periphery
have less connections perhaps are
connected to each other which hub Aegina
is connected to can tell us a lot about
its function etc we can learn a lot
studying the social network of genes
like on Facebook when I find out who
your friends are for example if you have
friends who University students and like
I don’t know climbing mountains then
you’re probably a student who likes
climbing and cetera so in this again
when I see who are the Friends of a
given gene I can make a very good
prediction about what this gene does but
this is just the beginning we need a
more deeper way of describing it and of
course any deeper knowledge in science
comes from mathematics and mathematics
provides us with a formal way of
describing networks the seminal paper
was published in the 1990s by a
Hungarian American mathematician Albert
laszlo barabasi and he described a way
of predicting network behavior based on
their properties and the cool thing that
the mathematics is not a terribly hard
I’m not a mathematician I’m a biologist
and I can follow it if you are okay with
high school level mathematics you can
follow it and the handbook is available
online it’s open and you can all try and
look how it how it works and how it can
be applied to biology to social networks
to politics to economics to anything but
is that enough is describing those
networks enough where we are at now in
genetics is essentially making
inventories of parts we take the DNA of
different organisms we list the genes
that are there and by a very slow and
expensive process we find out the
interactions between them we make those
maps we may we map all those connections
but this even this may not be enough
because these are just connections
between two elements what if they’re
real true knowledge is hidden in the
interactions that involve three or four
elements at the same time then you draw
such map not on a two-dimensional plane
but in an n-dimensional space that’s
where it gets harder even for
mathematicians so maybe we need
something more something more
maybe not human because in modern
science we realized that artificial
intelligence algorithms are very good at
spotting regularities in enormous sets
of data all this very famous big data I
thought this is what’s used by companies
that gather all this seemingly
irrelevant information about us combine
it into huge databases and then use it
to make more money of us using machine
learning using computer programs that
learn to spot patterns in huge oceans of
data so maybe this is the approach that
can help us in genetics and it’s been
tried recently in September last year
there was a paper where the scientists
took faces of a few hundred people
measured them and fed this data into the
computer machine learning program and
then gave the same program the
information about the complete genome
complete DNA sequence of all those
person
and made this artificial neural network
learn the patterns and the program
learnt to predict how a person’s face
looks like based on the DNA on the left
is the photograph on the right is what
the program gives you so as you can see
they kind of look like this there is
there is a huge discussion about this
paper whether it’s really correct or not
whether it’s really very predictive or
not if they were the replies counter
replies etcetera but this is the
beginning of something that can be very
promising that can also be the end of
science
why because science is about explaining
the complex by something which is
simpler about this is how our
understanding of the world increases and
here we have a machine that makes the
prediction it gives us the practical
aspect of science it works it can
predict something but we don’t
understand any more than we did before
it doesn’t increase our knowledge
because what happens in this algorithm
is as complex as s and as impenetrable
to us as the biological reality so if
this is the only way this is the end of
science as we know it I hope not but
we’ll see it’s going to be exciting
either way we travel way it progresses
and we have only started to understand
how complex the problems of genetics are
one thing is certain and it was already
noticed by a great scientist a great
physicist who at the beginning of our
century said I think the next century
will be the century of complexity and
this is true in all different branches
of science genetics included and this is
why genetics is still hard because we
know the parts we know the elements we
know the same single leg
bricks we know how they work but now we
need to understand how they interact and
how they combine to make complex system
and this is the future of genetics in
the next decades and for that we need
one important thing we need to be
interdisciplinary because we need
collaboration between biologists
mathematicians physicists philosophers
maybe even humanists so what is really
important is what you did during the
break networking where biology students
networked with mathematics or physics or
computer science students of philosophy
students this is necessary to build the
future of science I’m not saying we
should turn down the walls of the
physics building or the biology building
because we’d freeze if we did that but
we need to do it metaphorically we need
to tear down the walls between different
disciplines that exists in our minds and
that’s my take-home message for today [Applause]
a half has been called the century of
biology Elevens essentially of genetics
because it’s really been an fantastic
and interesting adventure because as
recently as in the middle of the 19th
century we had no idea how heredity
worked
starting with Mendel’s work and that
actually entered the scientific discuss
only at the beginning of the 20th
century we started with the idea that
genes are some abstract entities that
determine certain traits of organisms
for example there is a gene that makes
plant grow tall or short there is a gene
that makes the flowers red or white
there is a gene that makes flies ice
white instead of red etc etc so abstract
entities that determine certain traits
and from that in about a century we went
to this to a very good understanding of
the very chemical and physical language
of the genes we know that genes are made
of DNA we understand the alphabet of DNA
we can read it we can even rewrite it we
can take a gene from one organism and
put it in a different organism like here
a gene format relative was put into
different animals making them
fluorescent by the way this is not just
a party trick this is a very good tool
to study where genes work in a given
organism and soon some say we might get
to modifying humans to getting
custom-designed
babies with better mine’s better bodies
better everything well not so fast it’s
not as easy
as it seems even though thanks to this
enormous development of genetics from
nothing at the beginning of the 20th
century to where we are now we know
pretty well how a gene works we
understand how ATC’s and G’s get read
gets translated into proteins that make
our cells biologic textbooks are full of
this information of course we still work
on the details but we already have a
pretty good idea about how it works
these days I could take DNA from any one
of you and treat every single letter
written in this DNA it would take a few
days and about a few hundred cures to do
it for any single individual and the
databases are full of such information I
can find genes in the DNA of any other
organism although if you look carefully
at those gene lists you will mostly see
the names of awful diseases that happens
when one of those genes goes wrong you
don’t see a gene for intelligence beauty
talent etc why is that why is that not
so simple because maybe if we can read
DNA if we understand DNA then maybe we
can realize what has for a long time
being a sci-fi vision that we can take
DNA from anyone and predict their future
their talents the possibilities that
determine the entire future life well
not so fast
let’s start with something simple
something that may be familiar from
those of you who took some genetics in
high school there is not a single gene
that determines the color of your eyes
if you believe that from your school
then it’s wrong but it’s still simple
there are about six genes that make your
eyes dark brown or bright blue when you
read and check about a dozen genes you
can make a pretty good prediction about
a person’s color of her or eyes again
not as simple as it might seem but
doable you need a dozen genes some
mathematics and you can predict whether
some of somebody’s eyes look like this
or like this but that’s still a simple
example let’s take something more
challenging how tall you are there is
not a single gene that makes you tall or
short to understand why some people grow
real tall and some not so much you need
to look into at least 200 different
genes and I can tell you that at the
moment we don’t have enough mathematical
theory of how those genes interact to
make this prediction I cannot take your
DNA and predict all you should be about
6 feet tall no it is not possible it can
be done for 4 traits that depend on a
dozen genes but not on those that depend
on 200 genes and that’s all not yet that
complex because if we look at some other
traits
is there a gene that makes you behave in
a certain way for example is there a
gene that makes you more likely to enjoy
dangerous behaviors like jumping from a
perfectly good airplane there were
headlines in the newspapers that there
is a gene that makes you a risk-taker
no it’s not true complex behaviors
depend on many genes there is not a
single risk-taking gene there is not a
single intelligence genes there is no
smart gene at least a thousand or two
thousand genes would be necessary to
explain the hereditary basics of of
intelligence there is no not a single
gene
that makes you susceptible to common
diseases like heart diseases like most
cancers like autoimmune diseases etc
there are some diseases that depend on
single genes but they are very very rare
the common ones depends on thousands of
different genes so the common theme is
that you cannot explain the traits of
organisms by looking at single genes you
need to look at the interactions between
many genes and that’s a very common
theme in the entire field of biology
that complexity is built from
interaction of simple elements like
those Lego bricks it’s not hard to
understand a single element those bricks
children understands them but the
complexity lies in the interactions
between those elements human genome has
about 20,000 genes so 20,000 bricks and
the complexity of this is not the
complexity of the single elements it’s
their combinations it’s their
interactions so in need if you want to
understand the whole you need to
understand not just the elements but how
they interact and how can we approach
those complex interactions in genetics
one of the ways to look at it is to
realize that interacting elements form
networks and the world is made of
networks here you can see airline
connection networks anybody who’s
planned a longer journey has experienced
this you can see that these are not just
random connections there are hubs that
are connected to each other and
peripheries that are connected to hubs
and this can be described in a certain
formal manner your brain is also a
network it’s composed of a hundred
billion elements called neurons and we
start to map all those interactions
there is a project so
human connectome project to understand
all those connections and the elements
are quite simple they are called neurons
or nerve cells and we understand them
already pretty well we can model how a
single nerve cell behaves but we take a
hundred billion of them and we have your
mind unpredictable and complex the
complexity is built on interaction and
knowing the elements is not enough to
know the entire system another example
social networks people interact with
each other and these days everybody does
it online that’s an example of a Twitter
followers Network that’s Facebook
connections and if you look at those
this picture it may resemble what you’ve
seen on this airline connection map
there are also cups and peripheries
everybody has this friend who knows
thousands of people and when you see who
is friends with those central hubs you
can find out that few steps are enough
to connect you to about anyone there is
a theory that says that six handshakes
six steps are enough to connect any two
people in the entire world if the world
is made of networks we need some means
some way to understand to describe
networks and we can do it in biology as
well so let’s make a Facebook of Dean’s
let’s make a social network of genes
that’s a picture from a recent paper
showing interactions between genes in a
simple yeast cell as you can see the
scientists map mapped all the
interactions it had on it has only about
6,000 genes so less than a human cell
and you can see that there are genes
that are hubs that have many connections
and there are periphery
have less connections perhaps are
connected to each other which hub Aegina
is connected to can tell us a lot about
its function etc we can learn a lot
studying the social network of genes
like on Facebook when I find out who
your friends are for example if you have
friends who University students and like
I don’t know climbing mountains then
you’re probably a student who likes
climbing and cetera so in this again
when I see who are the Friends of a
given gene I can make a very good
prediction about what this gene does but
this is just the beginning we need a
more deeper way of describing it and of
course any deeper knowledge in science
comes from mathematics and mathematics
provides us with a formal way of
describing networks the seminal paper
was published in the 1990s by a
Hungarian American mathematician Albert
laszlo barabasi and he described a way
of predicting network behavior based on
their properties and the cool thing that
the mathematics is not a terribly hard
I’m not a mathematician I’m a biologist
and I can follow it if you are okay with
high school level mathematics you can
follow it and the handbook is available
online it’s open and you can all try and
look how it how it works and how it can
be applied to biology to social networks
to politics to economics to anything but
is that enough is describing those
networks enough where we are at now in
genetics is essentially making
inventories of parts we take the DNA of
different organisms we list the genes
that are there and by a very slow and
expensive process we find out the
interactions between them we make those
maps we may we map all those connections
but this even this may not be enough
because these are just connections
between two elements what if they’re
real true knowledge is hidden in the
interactions that involve three or four
elements at the same time then you draw
such map not on a two-dimensional plane
but in an n-dimensional space that’s
where it gets harder even for
mathematicians so maybe we need
something more something more
maybe not human because in modern
science we realized that artificial
intelligence algorithms are very good at
spotting regularities in enormous sets
of data all this very famous big data I
thought this is what’s used by companies
that gather all this seemingly
irrelevant information about us combine
it into huge databases and then use it
to make more money of us using machine
learning using computer programs that
learn to spot patterns in huge oceans of
data so maybe this is the approach that
can help us in genetics and it’s been
tried recently in September last year
there was a paper where the scientists
took faces of a few hundred people
measured them and fed this data into the
computer machine learning program and
then gave the same program the
information about the complete genome
complete DNA sequence of all those
person
and made this artificial neural network
learn the patterns and the program
learnt to predict how a person’s face
looks like based on the DNA on the left
is the photograph on the right is what
the program gives you so as you can see
they kind of look like this there is
there is a huge discussion about this
paper whether it’s really correct or not
whether it’s really very predictive or
not if they were the replies counter
replies etcetera but this is the
beginning of something that can be very
promising that can also be the end of
science
why because science is about explaining
the complex by something which is
simpler about this is how our
understanding of the world increases and
here we have a machine that makes the
prediction it gives us the practical
aspect of science it works it can
predict something but we don’t
understand any more than we did before
it doesn’t increase our knowledge
because what happens in this algorithm
is as complex as s and as impenetrable
to us as the biological reality so if
this is the only way this is the end of
science as we know it I hope not but
we’ll see it’s going to be exciting
either way we travel way it progresses
and we have only started to understand
how complex the problems of genetics are
one thing is certain and it was already
noticed by a great scientist a great
physicist who at the beginning of our
century said I think the next century
will be the century of complexity and
this is true in all different branches
of science genetics included and this is
why genetics is still hard because we
know the parts we know the elements we
know the same single leg
bricks we know how they work but now we
need to understand how they interact and
how they combine to make complex system
and this is the future of genetics in
the next decades and for that we need
one important thing we need to be
interdisciplinary because we need
collaboration between biologists
mathematicians physicists philosophers
maybe even humanists so what is really
important is what you did during the
break networking where biology students
networked with mathematics or physics or
computer science students of philosophy
students this is necessary to build the
future of science I’m not saying we
should turn down the walls of the
physics building or the biology building
because we’d freeze if we did that but
we need to do it metaphorically we need
to tear down the walls between different
disciplines that exists in our minds and
that’s my take-home message for today [Applause]
Please follow and like us: