Player is loading...

Embed

Copy embed code

Transcriptions

Note: this content has been automatically generated.
00:00:00
Uh after finding out last night many of
00:00:02
you from or nutrition background. And
00:00:04
less of the market by background to
00:00:06
explain a little bit of the history
00:00:08
have some of the inscrutable diagrams
00:00:10
saying at the last year as as I I've
00:00:12
lost a site. And also some of the how
00:00:15
we developed Some of the techniques
00:00:17
lead to those so begin with that then
00:00:19
talk about how we apply them problems
00:00:21
of AB this remote users so the
00:00:24
backgrounds well the verses as you IDNA
00:00:26
sequencing getting dramatically cheaper
00:00:28
so much so that would be quite
00:00:30
sequences energy lyman two thousand
00:00:31
nine it was literally almost a million
00:00:34
times cheaper than the for the beam
00:00:35
into thousand one when the human genome
00:00:37
project was announced complete and it's
00:00:39
got it's got a couple bottles to make
00:00:41
the CT something. But then the problem
00:00:43
is you have to interpret all of that
00:00:45
sequence data so what I'm saying here
00:00:47
is just nine out of a hundred thirty
00:00:49
thousand sequences project we did with
00:00:51
the first lap about the reason like
00:00:53
this is what happens if you use a file
00:00:55
containing the final genetic tree the
00:00:56
related sequences to each other. And
00:00:58
that's what happens if you just
00:01:00
visualise the tree. So you can see a me
00:01:02
really that you have your work cut out
00:01:03
for you interpreting all the sequence a
00:01:05
local street data. And that doesn't
00:01:07
even help me that much I tell you study
00:01:09
design for this particular project what
00:01:11
we were doing this by geography but not
00:01:13
on the grand scale and one Wallace have
00:01:15
the nineteenth century spores you match
00:01:17
the distribution of logical organisms
00:01:19
across the we was study my card so we
00:01:22
thought would be able to find by
00:01:23
geographical patents much places time.
00:01:25
So what we did was we took three
00:01:26
keyboards and then I think it's it's of
00:01:29
the people we typed a small all the
00:01:31
keys not think is hands and initially
00:01:33
we just wanted to answer basic
00:01:34
questions by geography like the
00:01:36
spacebar have more markets are not
00:01:37
simply because a lot of the mail but
00:01:39
he's that is that is that with you my
00:01:41
could survive it lush valleys your the
00:01:43
prince. And so it was here by
00:01:45
geographical dividing line between say
00:01:47
the GQBHP because the the right the
00:01:50
right next to each other on the
00:01:51
keyboards as you type in there of any
00:01:53
type of opposite hands. So depending on
00:01:55
how the space to what you might see
00:01:57
different patterns. So that the the
00:01:59
data registry idea makes the main thing
00:02:01
going on on the system immediately
00:02:03
clear so on this diagram each point
00:02:06
represents. all the markers and on time
00:02:08
I heard look you know she identified by
00:02:10
the DNA And so it's points comes from
00:02:12
either negative or from a key on the
00:02:14
keyboard defined colour them by paced
00:02:16
and you see the points this person of
00:02:18
the second person and that that this a
00:02:20
cluster together or some contrast if I
00:02:22
call them according to what he but I
00:02:24
think it you see my separation. So you
00:02:27
can tell immediately that we each have
00:02:28
a unique personal skin community we
00:02:30
transferred from I think it's it's like
00:02:32
people to suite at another schematic
00:02:34
PNES few LC but more importantly labels
00:02:37
on the TVCCSI Miami so you really know
00:02:39
it straight so so fine that aside so
00:02:44
also I'll tell you a little better
00:02:45
that's we produce these diagrams what
00:02:47
was also principal coordinates plus you
00:02:50
different distances well explain what
00:02:51
lays me later I one one one key problem
00:02:55
with be applying to you as a beastly
00:02:57
which is a math problem globally not
00:03:00
just in the united states but
00:03:01
increasingly in the rest of the world.
00:03:04
And because there are many different
00:03:05
actively contribute to to ABC
00:03:08
everything from psychology to die and
00:03:10
so fourth. And there are many human
00:03:12
genes that contribute to obesity this
00:03:14
is just the buttons just a recent
00:03:16
review assuming some of the ACL say
00:03:18
documents that we now know contribute
00:03:21
to obesity. But the problem is that the
00:03:23
project and the value of these generic
00:03:25
models a very porous so if you take
00:03:27
every humans that ever identified by D
00:03:30
bosses contributing to it easy you can
00:03:32
only predict easily and he's a piece
00:03:34
with about fifty eight detectors which
00:03:36
is not very impressive say says a lot
00:03:41
of work that I'm gonna sell you balance
00:03:42
in collaboration with yeah and about
00:03:45
about a decade ago he became one of the
00:03:47
leading I don't suppose hypothesis that
00:03:49
the obesity epidemic might actually be
00:03:51
an epidemic in terms of the transmits
00:03:53
possible agent. And the first what's
00:03:55
really but between the these
00:03:56
individuals to different the market all
00:03:58
communities you've already seen some
00:04:00
data from disk a and from Frederick
00:04:02
other person straight but yesterday we
00:04:06
would expect change weights to Carly
00:04:07
changes microbial communities and we
00:04:10
will also explain altering the
00:04:11
microbial communities should also
00:04:13
weight gain they lost. Um and say what
00:04:16
what I'm gonna show you here building
00:04:18
on what you saw yesterday is that all
00:04:20
three of these two humans sorry all
00:04:23
three of these attribute mice buffet
00:04:25
straight in the face to intrude humans
00:04:27
amalgamated helen's at the moment is to
00:04:29
figure out how we do that. and humans
00:04:30
as well. Um solicitous of this problem
00:04:34
of how we measure at least microbial
00:04:35
communities which are very calm place
00:04:37
again as you ideas today there millions
00:04:39
of jeans and we got hundreds of
00:04:40
difference P C.'s and you can just slap
00:04:42
annotate major figure out you can you
00:04:46
like. Um and in this context maybe but
00:04:50
still thinking be colours a classic got
00:04:51
my great but the reason why we think
00:04:53
that is not the so that is already to
00:04:55
the entire again me and again pretty a
00:04:57
bacterial in the file and as a major
00:04:59
player and they got microbe I'm it's
00:05:01
just that you call I was really get a
00:05:02
living in captivity unlike the vast
00:05:04
majority of my kids that are out there
00:05:06
in the environment and that are in
00:05:08
there Larry got and say figuring out I
00:05:10
have to deal with this and culture
00:05:12
majority involves a lot of the way you
00:05:14
hold that that you had about yesterday
00:05:16
and a lot but you will hear about
00:05:17
today. Um say so so you a lot of one
00:05:21
and gives tell you about today is gonna
00:05:23
be based on studies of the rubber
00:05:24
similarity G other tell you about that
00:05:27
X number infantry which pieces of
00:05:29
presents we got I I believed in long is
00:05:32
get a tell you much more about change
00:05:34
management mix we yes where where you
00:05:36
get a view of the functions well and
00:05:38
what I'm getting like here is looking
00:05:40
at this P C.'s which is a very cheap
00:05:41
way of doing these studies can tell you
00:05:43
a lot about about what you want to know
00:05:46
about different systems although it's
00:05:47
only one piece of the puzzle and
00:05:49
getting of the functions but much
00:05:51
recommended genetic sequencing event by
00:05:53
by the transcript I mix by the tablets
00:05:55
and so one is gonna be essential for
00:05:56
police station make innocence behind
00:05:58
many of these patents and show you But
00:06:01
the reason why we use rubber similar in
00:06:02
a is general P C.'s as part of the
00:06:05
writers and so also I hazard it has
00:06:08
fast evolving sly revolving regions so
00:06:10
we can we can truck phylogeny
00:06:12
relationships are different dates and
00:06:14
then there are these huge preexisting
00:06:15
databases we can Paris and so we don't
00:06:17
have to do everything from scratch from
00:06:19
scratch every time we can look out a
00:06:21
lot of us pieces that sample in these
00:06:23
existing databases so what we do direct
00:06:28
environmental sequencing to see the
00:06:29
vast majority of my cards will not be
00:06:31
able to go to the lab so the basic
00:06:33
principle here is to get the samples
00:06:35
extract the DNA so maybe maybe the to
00:06:38
comes from someone lean well we from
00:06:40
someone is that these in what I'm gonna
00:06:42
show you it's mostly based on PCR
00:06:44
amplification small subunit runs them
00:06:46
or lady of of course you can take all
00:06:48
the DNAD shotgun sequencing out as well
00:06:51
which you'll hear more about ways for
00:06:52
those what you also good bass several
00:06:55
I'm just today traditionally you take
00:06:57
the DNAG weak cleaners and you would
00:06:59
sequencers but the great thing about
00:07:01
agency. and saying as you can sequence
00:07:03
direct from the PCR product say every
00:07:05
time you run the high C we collect
00:07:07
about another but in sequences. So
00:07:09
we're saving a billions we expect
00:07:10
colonies abandon including files and a
00:07:13
fair amount of expense and there Um and
00:07:15
the traditional you take the sequences
00:07:17
and blast immigrate them according to
00:07:18
what was the motive them and inbound
00:07:20
but ironically it's a success this
00:07:22
culture in the you know methods that
00:07:23
makes at least less useful because now
00:07:25
instead of having and instead of having
00:07:28
be like them or something else what the
00:07:29
name of the phenotype you recognise you
00:07:31
get ahead and cultured environmental I
00:07:33
slept number thirty eight minutes seven
00:07:35
one of these studies. So what we have
00:07:37
to do it is taken explicitly pilot
00:07:39
genetic approach where you do multiple
00:07:41
sequence alignment bodybuilder tree and
00:07:44
then you relational sequences you found
00:07:45
that our genetic tree to the
00:07:47
environmental samples looked at but
00:07:50
that she based approach has its own
00:07:52
problems so this picture is a pretty
00:07:54
hard to understand the to analyse what
00:07:56
I'm saying here is data from we did
00:07:59
with the ones that back in two thousand
00:08:01
five looking at five thousand sequences
00:08:03
that the mouse got and twelve thousand
00:08:05
sequences a human column from their
00:08:06
problems like in red and blue ones
00:08:09
respectively and at the time this was a
00:08:11
huge number of C concerns about now
00:08:13
surrounding your for sample what you
00:08:15
get back from the sequence but what you
00:08:17
can see there's a bunch of red clusters
00:08:19
in the budget clusters at the time we
00:08:21
did this there was significant
00:08:23
statistically tells the reasonably
00:08:24
distributions will not identical to
00:08:26
each other. And that might not seem so
00:08:28
that on the street what what two
00:08:29
colours but I haven't ODS as that
00:08:31
actually we were looking at twenty one
00:08:33
different even samples and samples from
00:08:35
nineteen different mice but that I we
00:08:36
paper battery forty different colours
00:08:39
you try to figure out a sigh and most
00:08:41
similar so a well also a three looking
00:08:43
at a buyer I you really have your work
00:08:45
cut out and just illustrate the problem
00:08:48
the goal of this is not a company's
00:08:50
particulars as just the paper that was
00:08:52
at the top of my stack when I was
00:08:53
together what we're doing is very
00:08:55
valuable right there looking new right
00:08:57
we got out him I god they could be
00:08:59
involved in like this endless
00:09:00
degradation by feels about the issue is
00:09:03
you Reading through the paper you come
00:09:04
to think it through it's pilot genetic
00:09:06
tree you sequences safe and had this
00:09:08
little short labels the longer
00:09:10
sequences in the database of these
00:09:12
longer labels fixation numbers and so
00:09:14
you can say whether new sequences
00:09:15
better that's a very well. But then you
00:09:17
come to figure for which is another
00:09:19
tree forget five and think a six and a
00:09:22
whole lot of additional thing is what
00:09:24
the tragic and say what we had to do
00:09:29
was we had to get out of that first of
00:09:31
individual is inside capitalise a pain
00:09:33
really that's even to my lap you know
00:09:36
has an independent faculty position of
00:09:37
Denver the the put together this really
00:09:41
nice H decreases based on the side of
00:09:43
calculating a community distance metric
00:09:45
based on evolutionary history. And so
00:09:48
the idea is very simple so the idea is
00:09:50
that if you have unrelated communities
00:09:52
all of the branches on the tree again
00:09:54
to leave only to members of the red
00:09:55
community or only yeah the bleep
00:09:57
community but not about right because
00:10:00
you can have something from one
00:10:01
community moving to the other and
00:10:03
surviving reproducing in continuing to
00:10:05
add gradually. So that communities
00:10:07
around releases we define that just
00:10:09
there's one there's far out as they can
00:10:11
possibly be wasn't contrast if we have
00:10:13
two identical communities with complete
00:10:15
something every my could you find of
00:10:17
the red community you find blue. So all
00:10:20
the branch think much reshaped by by
00:10:22
communities should your paypal and we
00:10:24
define that distances era. And then if
00:10:26
you have two related communities just
00:10:28
see some shared branches paypal and
00:10:30
some unique ones available we we define
00:10:33
the unique fraction or you know
00:10:34
practise distance on the tree as a
00:10:36
fraction of the tree that leads at
00:10:37
least one community rubber leading to
00:10:39
but so what's so great about what's
00:10:42
well what we can do is we can use that
00:10:44
one's inside that there is one tree of
00:10:46
life makes all cousins to take my could
00:10:49
you any environment item on the same
00:10:51
tree so here we have one tree with the
00:10:53
red yellow and blue microbial
00:10:55
communities coloured. And say we can
00:10:57
compare using that unified technique I
00:10:59
just showed you red yellow relatively
00:11:02
relatively summarise all of these
00:11:04
distances in the distance matrix. And
00:11:06
then once we have that distance matrix
00:11:07
we can use techniques I principal
00:11:09
components analysis hierarchical
00:11:11
clustering another mo. right
00:11:13
statistical she needs to relate all the
00:11:14
communities one another at the whole
00:11:16
community level using the evolutionary
00:11:19
history the organisms that are involved
00:11:20
Um can spoke of is relatively similar
00:11:23
components analysis which many of you
00:11:26
have probably seen on the directory
00:11:27
context the differences instead of
00:11:29
starting with the actual what what the
00:11:31
actual locations and euclidean space
00:11:33
example what we start button's
00:11:35
distances no we construct a set of
00:11:36
principles that's all the distance
00:11:38
constraints and then we do
00:11:40
dimensionality reduction et cetera
00:11:41
points. So it's essentially like having
00:11:44
a having a flashlight where your
00:11:45
prediction shadows on the wall of the
00:11:47
points you data cloud really want to
00:11:48
raise a bachelor right right a that
00:11:50
fishing spare data points as much as
00:11:52
possible to explain as much of the
00:11:54
variability that strip this principle
00:11:56
access a menu right spaces at ninety
00:11:58
degrees to match to spread out points
00:12:00
as much as possible again that's just
00:12:01
taken principal axes and then you keep
00:12:03
on doing Nash explaining less and less
00:12:05
of the variation with each additional
00:12:07
access seven bicycle parts you saw
00:12:09
yesterday and that you'll see today
00:12:11
based on this technique of the success
00:12:13
of axes are explaining as much as
00:12:15
possible of the variability in the data
00:12:17
say other arbitrary coordinates you
00:12:19
derive intrinsically from the data
00:12:21
itself. So what does technique and and
00:12:24
camping initially decided she would
00:12:26
like us everything and say we went and
00:12:28
then put together this matter analysis
00:12:30
of a hundred studies two hundred
00:12:32
samples twenty thousand sequences
00:12:34
literally spending the will to
00:12:35
everything supposed to be there seem to
00:12:38
temperature PH just about every
00:12:40
imaginable fact to we could affect
00:12:42
microbial communities globally. And so
00:12:44
the question is out of all this
00:12:45
complexity just like all the complexity
00:12:47
that relates different people would
00:12:49
they be any clear patents that made
00:12:51
from the data. And much to my surprise
00:12:53
the answer was yes when you take all
00:12:55
these differences and always
00:12:56
environments you see this very this
00:12:58
last between the salient and than on
00:13:00
sale in environments use imagine a
00:13:02
hierarchical clustering as well as the
00:13:03
peak away this let's say station but
00:13:05
displays you'll less analysis piece
00:13:07
with live in the middle there what
00:13:09
those are mysteries where you truly
00:13:11
high. the mixture between the salient
00:13:12
but on sale in so we can catch of these
00:13:15
biological relationships among these
00:13:16
very heterogeneous environments. Um but
00:13:19
the cool thing about sort of really
00:13:21
matters how you measure these distances
00:13:23
you could people referring to do some
00:13:24
different distance matrix yesterday the
00:13:26
reason why you want to take evolution
00:13:28
into account when you mention based
00:13:29
distance Um this is the the slightly
00:13:32
updated you have that same data what
00:13:33
salient unsettling split the mysteries
00:13:35
in the matter with you know frank if
00:13:37
instead you take it able to access like
00:13:40
it's P C.'s level the genus level the
00:13:41
family level or whatever and you
00:13:44
customers data use of the distance
00:13:45
matrix like euclidean distances which
00:13:47
is very commonly used focus on
00:13:49
sequenced its metric measure or
00:13:51
whatever you get this spike artifacts
00:13:53
by two degrees relative. one another
00:13:54
because the data matrices very sparse
00:13:56
And you release things like the
00:13:58
biological significance of the S trees
00:14:00
begin to mediate as opposed to going
00:14:03
off in the right direction somewhere
00:14:04
and so the reason why is that when you
00:14:06
use a file genetically based distance
00:14:08
major you take the evolutionary history
00:14:10
into account when you don't do that
00:14:12
it's not you make no assumptions about
00:14:14
the rate you're actually you're
00:14:16
actually assuming that reassessed a
00:14:17
phylogeny where all be Caesar equally
00:14:20
related to one another and the matter
00:14:22
whatever seem may compile genetic
00:14:23
reconstruction whatever apology you
00:14:25
build is gonna be based on them at stop
00:14:27
phylogeny and explaining the biological
00:14:29
relationship and so you get
00:14:31
biologically more meaningful clustering
00:14:33
out anyway what one thing that
00:14:36
suppresses about this environmental
00:14:38
data it's not related to hasten to
00:14:40
humans just maybe one one thing that
00:14:42
the one thing that surprised us about
00:14:44
it was that extreme environments are
00:14:46
from our perspective without wires
00:14:47
wiper microbial perspective. So if you
00:14:50
high temperature environments like for
00:14:52
marine sediments from hot Springs yeah
00:14:54
we stayed but it's just like like the
00:14:56
other salient non so samples risk
00:14:58
actively about my colleague an
00:15:00
additional samples if ones that we're
00:15:02
able to find some really extreme
00:15:04
environments microbial perspective the
00:15:06
when we add them to the graph
00:15:07
introduces new access that explains
00:15:09
twice as much of the initial so silly
00:15:10
nonsense but I said it up. And and and
00:15:14
squash all these points from all over
00:15:16
the well just happen to one side was
00:15:18
graph. So you might be wondering how I
00:15:20
you have to get find those microbial
00:15:22
use extreme environments the answers
00:15:24
that you don't have to get started all
00:15:25
the right there with honest. say what
00:15:27
we're saying is mostly the million gas
00:15:29
as a clear our wireless part other
00:15:32
heist associated habitat site map
00:15:34
pictures here vagina not pictured here
00:15:36
and so forth are intermediate between
00:15:38
the two and then these these has to say
00:15:40
she's environments very different from
00:15:42
what we find free living or what we
00:15:44
find associated with them better
00:15:45
precise like sponges or are addressed
00:15:48
real better but So you save mice about
00:15:52
actually sector's been based on
00:15:54
traditional saying a sequencing and
00:15:56
what really allowed shifts to hire
00:15:58
people techniques with the observation
00:16:00
that we could just take a small
00:16:01
fragment the sixteen it G get the same
00:16:03
community clustering information as
00:16:05
what polling sequence reason why this
00:16:07
is really important is the the short we
00:16:09
technologies a much cheaper P sequence
00:16:12
other traditional thing the sequencing
00:16:13
all right now they're starting to catch
00:16:15
up with the reflectance anger. So this
00:16:17
is less but what we did was basically
00:16:19
took three datasets we we had me
00:16:21
appalling sequences from sixteen
00:16:23
described similar in a that's what
00:16:24
would happen if we only had a hundred
00:16:26
to two hundred fifty basis that
00:16:28
sequence but we get the same result or
00:16:30
different so what I'm saying here's
00:16:32
clustering of the humans and the bias
00:16:34
but I said you before elementary and
00:16:36
then they go raring agree microbial mat
00:16:38
which is a hyper silly green mask
00:16:39
that's the most this microbial
00:16:41
community we know about it and you can
00:16:43
say that despite that that this T the
00:16:44
humans and the buys a pretty different
00:16:46
from one another. So it's cut a long
00:16:48
story short if you take a good two
00:16:50
hundred base fragment you get exactly
00:16:52
the same clustering path right down to
00:16:53
the same as you can also take a very
00:16:56
bad two hundred base fragment from the
00:16:57
same they and in that case you would
00:16:59
make the amazing. every but some of the
00:17:01
humans almost similar to the mice but
00:17:04
they ask the other humans. So those
00:17:05
would be very exciting faced try you
00:17:07
get back to clinical data that's what
00:17:09
is it about these people that make
00:17:10
resemble the mice and have a vegetarian
00:17:13
till eleven restricted environment a
00:17:15
very tiny a well and then and then So
00:17:18
when you got a fully sequences and you
00:17:20
found out just wasn't sure you might be
00:17:22
disappointed but we get is also sixteen
00:17:25
assistant changing very rapidly
00:17:26
ascending uniquely tied substitutions
00:17:28
accumulating every minute noodles I and
00:17:30
so the eh regions are we getting two
00:17:32
thousand seven when we did this and the
00:17:34
same regions really get today and you
00:17:35
should just look at the region that's
00:17:37
the right wing for the technology you
00:17:39
use and so does it say my kind of the
00:17:43
this is no we can address by recap or a
00:17:46
say is now we met in maybe and number
00:17:49
of bubbles. in my lap but together the
00:17:51
software pipeline coach I'm sense a
00:17:53
qualitative insights and Mike really
00:17:55
colour Um it's free it's open source
00:17:58
run anywhere from the laptop to the
00:17:59
Amazon DCT class. So the hot CP
00:18:02
computer escort hundred forty thousand
00:18:03
calls just larger social run on to date
00:18:06
well I see there's alleged integrate
00:18:08
the analysis of hundreds of samples
00:18:10
simultaneously mostly we use that for
00:18:12
sixteen S and become sequencing you can
00:18:14
also use it for other marketers and you
00:18:16
can also use the shotgun method to
00:18:18
mimic scimitar transcript I'm it's now
00:18:20
almost right although those protocols
00:18:21
still experimental stage. So basically
00:18:24
what we do is we tag hundreds of
00:18:25
samples of DNA pockets we introduce we
00:18:28
make some together and initially we
00:18:30
sequence one four five four well they
00:18:32
more recently would switch to the
00:18:33
aluminium platform which answer so you
00:18:35
have to consider buttons maybe because
00:18:37
a sequence of the bucket we can tell
00:18:38
what sample came from so we could play
00:18:40
stop the multiple sequence alignment
00:18:43
build phylogeny is and the use of
00:18:44
powered least cost the data builders
00:18:47
principal K whatnot spots but you'll
00:18:48
see many of already about our shows you
00:18:50
can say you chase more say incidentally
00:18:54
closely coming from a nutrition
00:18:55
background you might be wondering vices
00:18:57
technology called for five for it ends
00:18:59
up that Fahrenheit for fifty four as
00:19:01
the temperature your money bins as soon
00:19:03
as you get into sequencing. So that's
00:19:05
getting cheaper all the time. But it's
00:19:07
still not free and we have to do a lot
00:19:09
of action every batch however have able
00:19:13
but face time we did this in the in the
00:19:15
two thousand save them so we we spent
00:19:17
twelve thousand dollars on a four five
00:19:19
four run and got back and and and got
00:19:21
back about a half a million sequences.
00:19:23
And set at the time of silence of the
00:19:25
second facility down the whole my lap
00:19:27
was still charging eight bucks to read.
00:19:29
So if we double there incident with the
00:19:31
new technology what it cost four
00:19:33
million dollars and still twelve
00:19:34
thousand and they wouldn't be finished
00:19:36
with the sequencing yes and what the
00:19:37
aluminium technologies stopped another
00:19:39
couple of orders of magnitude cost
00:19:41
since then say services listens
00:19:43
tremendous in terms of making more
00:19:45
research questions more accessible you
00:19:47
can literally forty thousand samples
00:19:49
now and it's not a big deal. Um say
00:19:53
celeste cost LA no no censors actors
00:19:57
lab but he's not here Nestle and well I
00:20:00
put together this initial map of the
00:20:02
human microbe I'm back in two thousand
00:20:04
nine and said but how do you market
00:20:06
combine studies up to this point is
00:20:08
people studying different parts of the
00:20:09
marker by might them out this skin this
00:20:12
is still that use different techniques
00:20:14
on different subjects on different body
00:20:16
side and figuring out what was easy
00:20:18
different people what was use different
00:20:19
body size what was used different
00:20:21
technologies was really hard. So what
00:20:23
we did is we rounded up run that seven
00:20:26
people and swap them out a swap them a
00:20:27
lot up to twenty seven locations on the
00:20:30
body which as you can imagine as a lot
00:20:31
of places to stick a teacher mistake in
00:20:33
the morning and what we see emerging
00:20:36
from is very clear path where the mouse
00:20:38
is very separate we got three separate
00:20:41
the skin as a lot more spread out as
00:20:43
their the habitat five nostrils
00:20:44
previously explored primarily by by was
00:20:47
the year and the here and so forth. And
00:20:50
so I give the same they said a colour
00:20:51
by different no different different
00:20:53
variable. you can say that sex has very
00:20:56
little effect well estimate is has very
00:20:58
little effect compared to the body side
00:21:00
with one it's So what about somebody
00:21:02
set up on the day something said by the
00:21:04
way I was on the antibiotics separable
00:21:05
a couple weeks ago for sinus infection
00:21:07
do you think that a matter I said most
00:21:09
late today Gary ethics committees we
00:21:11
could someone else was point so I
00:21:12
suppose we'll find out but this was
00:21:14
totally different not just got but also
00:21:17
skin and the out for everyone no study
00:21:20
of like three months later I know this
00:21:22
is what everyone else. So the study
00:21:24
before I was to look at everyone waited
00:21:26
a look at them again then wait three
00:21:28
months like at the beginning that way
00:21:29
today look again. So we could look at
00:21:31
different temporal scales as well. And
00:21:34
so what's cool about this stuff to give
00:21:36
us based overall picture variability
00:21:38
making them combine. I I don't expect
00:21:40
you to read the text on labels and
00:21:42
probably don't really care I just wanna
00:21:44
say that's a lot of left right symmetry
00:21:46
in the major Texas across the body. And
00:21:49
then here you you know frank distance
00:21:51
on the Y axis so a lot to bob people
00:21:53
are more different eighty samples from
00:21:55
the same body have habitats a bite them
00:21:57
out a less different two different
00:21:59
habitats button a habitat to from the
00:22:01
same this less different different
00:22:03
different people within a habitat
00:22:04
person to like today and pass a list
00:22:07
different collected three months apart
00:22:09
them and that was my staff site we
00:22:11
looked at the skin of the scantily
00:22:13
stable and then the guy was in that way
00:22:15
but else was kind of frustrating
00:22:17
because you really want to figure out
00:22:18
dynamics right so now study we liked it
00:22:20
for time points purpose and and the
00:22:22
human microbiology we spent a hundred
00:22:24
seventy three million dollars been ages
00:22:26
money looking at a maximum of three
00:22:28
time points papers. And the reason why
00:22:30
this is a problem so this is the
00:22:32
viewing software that we developed a
00:22:34
workable stuff so this a god with them
00:22:35
out and so on and if I wrote it as
00:22:38
around to give you an explicit access
00:22:39
the time you can you can see the three
00:22:42
month separation you can't really see
00:22:44
the one day separation on the scale the
00:22:46
kind of looks like some changes but
00:22:47
it's kind of but it's hard to quantify
00:22:49
the state is what you'd really like to
00:22:51
do is you really like to get back to
00:22:52
some of the same people in sequence and
00:22:54
every single day for like a year and a
00:22:56
half. So what the after they wouldn't
00:22:58
platform we could afford to do that and
00:23:00
it looks like this. And the concordance
00:23:03
between the two sequencing technology
00:23:04
is really nice right so you get very
00:23:06
comparable results when you do it
00:23:08
across different platforms if you use
00:23:10
the right it needs to collect the data.
00:23:12
And the great thing about that so that
00:23:13
we can actually starts to animation so
00:23:16
what I'm gonna show you is a six once
00:23:18
in the last two people the documents in
00:23:21
amoeba like points my have remanded
00:23:22
looking skin them out in this so if I
00:23:26
just I was scary you can immediately
00:23:27
see yourself just how terrible. yeah
00:23:29
there's I have constant the mountains
00:23:32
and have a gas into meeting And you can
00:23:35
just imagine have a sort of high
00:23:36
resolution tracking could be amazing
00:23:38
looking both healthy variation which is
00:23:40
what we see here and also looking at
00:23:42
people's responses to different
00:23:43
individuals whether the drug
00:23:45
interventions or nutritional or what
00:23:48
are the other thing that's interesting
00:23:49
here's what by the time we were
00:23:51
together and have all kinds of
00:23:52
opportunities to exchange markets with
00:23:54
one another you can see that we really
00:23:56
maintain as separate microbial identity
00:23:58
statistics one time that's mostly
00:24:00
around us so you can see a separation
00:24:03
but we have to market items I propose
00:24:05
period. So each of us has as individual
00:24:07
market by that's relatively stable at
00:24:09
the time okay so sit site and that's
00:24:17
what we disappointed where our markets
00:24:19
come problem the first place. And if
00:24:20
you have dogs orchid society you
00:24:22
probably have some dark suspicions
00:24:23
about that all but well which turned
00:24:26
out to be true by the way see just like
00:24:28
we can match you up to mouse whatever
00:24:29
ninety five percent accuracy we can
00:24:31
actually up to your dog by the market
00:24:33
easier. Um but all seriousness once
00:24:37
about what what happens immediately
00:24:38
after the most of what we did with
00:24:40
Maria accurate to make the smell like
00:24:41
an out in my you at at at at the
00:24:44
delivery might has a huge effect on the
00:24:46
initial. michael's so I think about the
00:24:48
regular way basically markets all of
00:24:50
your body what like models vegetable
00:24:52
community specifically worked in
00:24:54
contrast to deliver by C section all
00:24:56
your markets like like skin yes another
00:24:58
view that same day so with the bachelor
00:25:00
samples from the from all the models
00:25:02
and read all the body have chance to
00:25:04
actually do with the babies and paying.
00:25:06
And I hear all of the all of the skin
00:25:09
from the models and and that we all the
00:25:11
body habitat small see section babies
00:25:13
Michael And we think that this initial
00:25:15
inoculation might explain some of the
00:25:17
differences between batch release
00:25:18
deliver the see section babies that
00:25:20
again you cared about yesterday so I
00:25:22
would reiterate now we also look at how
00:25:26
the marker by undeveloped infants of
00:25:27
those reply looking at one of the one
00:25:31
kid I've that I'm starting with the
00:25:33
make Acadia and then getting at the two
00:25:34
and a half years supply so you have
00:25:36
what looks like a relatively smooth
00:25:38
progression ultimately resembling the
00:25:40
mobile sample what the steady increase
00:25:42
and I basically and service progression
00:25:44
looks pretty smooth must be better so
00:25:46
you another view of the later
00:25:47
integration with human market by
00:25:49
project data that might lead easy think
00:25:51
a little differently about it. And we
00:25:54
can also do this sort of thing cross
00:25:55
culturally so what it's called lab we
00:25:58
looked at we wanted for people in three
00:25:59
populations from so what I'm saying
00:26:01
here is just the kids is zero right to
00:26:03
eighteen years of why and for each
00:26:05
point we're looking at a distance
00:26:06
between that point the average of
00:26:08
adults most community. So what you see
00:26:10
as as profound change over the face
00:26:12
three years the black followed like
00:26:14
convergence more or less the adult
00:26:15
community by the time you have H three
00:26:18
analysis replicator an african
00:26:20
population read stuff American
00:26:22
population green then the US population
00:26:24
and blue. very more time courses.
00:26:27
However way you wind up depends very
00:26:30
much more population you Um so if you
00:26:32
do a principal components part of the
00:26:33
same data puzzles coloured by age for
00:26:35
the young kids out of the adults and
00:26:38
like the different populations you can
00:26:40
see the US population as profoundly
00:26:42
different from allowing the under and
00:26:44
the amber indians. And of the bible and
00:26:46
we don't know how much of this is due
00:26:47
to die how much to genetics how much
00:26:49
environmental exposure because all of
00:26:51
these things very in different
00:26:52
populations. And that's context and
00:26:55
it's important to remember the project
00:26:56
slightly to they and also like ms
00:26:59
european cable and other very important
00:27:01
but there and exploring this much like
00:27:03
this to be looking at looking at least
00:27:05
an adult and the same as you go to
00:27:06
other populations and the scene is you
00:27:08
going to kids you find all these I
00:27:10
provide configurations you just don't
00:27:12
see matt initial population and that's
00:27:14
really really important for for the
00:27:16
drawing conclusions whether it's about
00:27:18
drugs or whether it's about nutrition
00:27:20
in one population and then trying to
00:27:22
apply these conclusions to a completely
00:27:24
different population what the
00:27:25
completely different marker by now that
00:27:27
we know that might provide has all
00:27:29
these effects on nutrition that you had
00:27:31
about yesterday and the you'll so hear
00:27:33
about today so was that have of the
00:27:36
rest of the database to buy a bike
00:27:37
happening with that they all American
00:27:39
got up with the idea was that we could
00:27:41
use the drop and sequencing tighten up
00:27:43
the ability to participate in this
00:27:45
research to anybody who was interested
00:27:47
and so you wouldn't pay a hundred
00:27:48
million dollars to find out what was
00:27:49
and you got with the drop cost and
00:27:51
sequencing technology we can do it for
00:27:53
a donation of about a hundred dollars.
00:27:55
So that's point will probably raised
00:27:57
about a million dollars for members of
00:27:58
the public almost all about ninety nine
00:28:00
dollar increments for people interested
00:28:02
in putting themselves lament microbial
00:28:03
mat and say and services and a Mcdonald
00:28:07
handing out one of the active biology
00:28:09
student smile at a hunting acted so the
00:28:11
person would you in project meeting
00:28:13
last year where are we have four
00:28:15
hundred people you have an if humans
00:28:17
you don't see as well as having them I
00:28:19
combine sequence so we can look at
00:28:20
relationships between them and we we
00:28:23
have a paper with replaying to inspect
00:28:25
a command excel in a couple weeks not
00:28:27
almost population but on a related what
00:28:29
looking at some of the terrible
00:28:30
components of the market by the
00:28:33
demographics lettuce cover a lot more
00:28:34
range than previous studies say upset
00:28:37
for example relative to the HMP we have
00:28:39
a lot more older people and we have
00:28:40
younger people we also have a lot more
00:28:43
obese people about the way people who
00:28:45
were primarily excluded from the HMP on
00:28:47
the grounds of their health conditions.
00:28:49
Um but the overall result we get a very
00:28:51
consistent what what the study's so
00:28:54
looking at the distribution of the P
00:28:55
C.s the the are samples of skin samples
00:28:58
comparing what we if self collected for
00:29:00
members of the public male than room
00:29:02
temperature this is what was clinically
00:29:04
collective immediately my case analysis
00:29:08
which yes and talk a little about how
00:29:10
the market by relates to genetics and
00:29:11
health. And one thing that was very
00:29:13
surprising to us services wait what
00:29:15
this lab and two thousand nine looking
00:29:18
one is a got it this is dies I collect
00:29:19
twins the relationship with a basically
00:29:21
so when we like them on as a got twins
00:29:23
on average the community's not
00:29:25
statistically significantly more
00:29:26
similar up and dies I got twins and
00:29:30
then and then looking S what
00:29:32
relationship to family members of more
00:29:33
similar including the files as we
00:29:35
should not more study but no yeah it's
00:29:37
an ink about all twenty twelve one and
00:29:39
then the family study the dogs I told
00:29:41
you about but unrelated it does appear
00:29:43
there away. So on the face but also
00:29:46
tell you that was no you genetic
00:29:47
effects on the microbe I'm but that's
00:29:49
very difficult to reconcile with what
00:29:50
we know but mice so for example in the
00:29:53
the and and the the the what uses which
00:29:57
which rhetoric back it was one of the
00:29:59
collaborators novels slip by with laid
00:30:01
out "'cause" now I see this profound
00:30:03
alteration of the market by in the
00:30:05
eyepiece mice oh but then you have this
00:30:07
problem with establishing causality you
00:30:09
don't have to take community as course
00:30:11
or result we at least. Um then what
00:30:15
what then with his group at Emory look
00:30:17
to the different system that's yellow
00:30:18
five knockout mice. So again they
00:30:20
become obese relative to wild type
00:30:22
mouse I again this profoundly also
00:30:24
microbial community between the T laugh
00:30:26
part knockouts about type correlated
00:30:28
with high I had as their track that's
00:30:31
right in cholesterol and even high
00:30:33
blood pressure. And test originally
00:30:35
fascination lee what's going on here
00:30:37
you can even transmit this ABC to
00:30:39
another mouse by transmitting the also
00:30:41
microbial communities so you have a
00:30:43
mutation that triggers this change
00:30:44
microbial community but then when you
00:30:46
transmit that microbial community to
00:30:48
genetically normal mouse is being
00:30:49
raised in free but now I'm what they
00:30:51
market outside that mouse also become a
00:30:54
beast the fascinating thing with the
00:30:57
system is the reason why they become a
00:30:58
base. So it turns out the limelight
00:31:00
BIBIB model where it seems like that's
00:31:02
the difference energy have us but the
00:31:04
elephant knockouts what you're saying
00:31:05
is but what what you're seeing is
00:31:08
primarily a difference in behaviour
00:31:11
somebody a five block at each more than
00:31:13
about type mouse that's I did become a
00:31:15
beast and similarly the gym pretty my
00:31:17
so the genetically normal again was
00:31:19
also microbial community also become
00:31:21
hungrier also libraries and also become
00:31:24
a these so it's a behavioural
00:31:25
phenotype. We had no idea how this
00:31:28
works I used was impossible mechanisms
00:31:30
and in in in Frederick stroke yesterday
00:31:32
but another that and we have the
00:31:33
scouting drawn just remind you that you
00:31:35
got my critics massively outnumber your
00:31:37
own cells in your body maybe a poison
00:31:40
you when it comes time to make choices
00:31:41
cafeteria one thing that's interesting
00:31:45
is the the mice diet has a much larger
00:31:47
effect and you decide so what the
00:31:49
decreases great last year what we did
00:31:52
was we put fifty two back strains of
00:31:53
mice on high that that simply this is
00:31:56
like that that's great you can save the
00:31:58
directory fate is much larger than the
00:32:00
genus activate. Um and you might be
00:32:02
wondering if any of this since humans
00:32:04
so what the with the the winter wasn't
00:32:07
yes that what they do they put people
00:32:09
on a die for year what resource changes
00:32:12
microbial community that led them to
00:32:14
resemble Colleen community and that
00:32:16
regularly the more individual lost the
00:32:18
big at the change in the microbial
00:32:20
community. Um we we are ten nights it
00:32:23
was a very talented that's to my lab is
00:32:26
yeah I remember the administrator was
00:32:28
also able to shave in that in the next
00:32:31
two thousand nine at dataset other we
00:32:33
can classify people we were peaceful
00:32:35
ninety percent accuracy based on only
00:32:37
got combine and this goes also showed
00:32:39
most certainly in the micro be
00:32:41
submission populations but you get very
00:32:43
good classifier accuracy for at least
00:32:45
now that's why not be very impressive
00:32:49
to the commercial odious previously you
00:32:51
this probably easier ways to tell if
00:32:53
someone's obese right and sequencing
00:32:54
other migrates but on the other hand
00:32:57
remember we can only do this with fifty
00:32:58
eight percent accuracy based on human
00:33:00
genes this is ninety percent accuracy
00:33:02
based on the market by on that and then
00:33:06
and then also in humans what we see is
00:33:08
a long tail but not sure to entire
00:33:10
correlates with major several patents
00:33:12
the market combine especially the mouse
00:33:14
a crap as hell and back to reality said
00:33:16
this is what what care it was great
00:33:18
okay however I however we needed
00:33:21
intervention study of the short too
00:33:23
what we find is that individuals
00:33:25
clusters so whether we put these it's a
00:33:27
people points colours coloured by
00:33:29
individual here these are people who
00:33:31
worked in place of a low fat diet or
00:33:32
hike at that what you can see is all
00:33:35
the points the same person class
00:33:36
together. And say are clearly longterm
00:33:39
dietary changes that be needed to make
00:33:41
these large changes in my by one other
00:33:44
thing I should point out is that by
00:33:46
mark as are problematic widespread
00:33:47
basically I end for diabetes level
00:33:49
studies have been done across nations
00:33:51
we have a paper that just came out its
00:33:53
little services really hot off the
00:33:55
press china for IBD you have a
00:33:57
consistent signal across different
00:33:59
populations by liability in terms of
00:34:01
the markets are involved we did not see
00:34:03
that in the different that what sort
00:34:05
been studied for at least three say a
00:34:07
sizable there was somehow what's best
00:34:09
as difference in the back tried it is
00:34:10
to immediate ratio that is not
00:34:12
generally tree and we need to look for
00:34:14
more subtle by Marcus probably high
00:34:16
level ecological patents relation
00:34:18
different people. So you might be
00:34:20
wondering can we do anything about that
00:34:22
actually changed my combine to of
00:34:23
disease but we going what this is
00:34:26
trying to develop kinda this last made
00:34:28
some based not on very subtle
00:34:29
differences in the human genome based
00:34:31
on much more obvious differences in the
00:34:33
market by "'em" a we can deploy even
00:34:34
and about in countries it's a tiny have
00:34:37
to K then then you guys to decide out
00:34:40
second you know like the spider
00:34:42
malnutrition study site what what the
00:34:44
Malawi negates project where and and
00:34:47
the you can see for example that
00:34:49
despite the fact that we discussed
00:34:50
population they are able to afford cell
00:34:53
phones. And what's fascinating about
00:34:54
the sincere now I have a five billion
00:34:56
active cell phones on it which is
00:34:58
pretty impressive when you consider the
00:34:59
only seven but people up the balloon
00:35:02
force people in the world twenty
00:35:03
percent of them have their itself and
00:35:05
and the one the reason for this most
00:35:07
floor it's a cheap set up digital
00:35:09
signal processing because of the
00:35:10
decrease in computation cost but it's
00:35:12
effectively break but stay not as to
00:35:15
allowing that for example I take the
00:35:16
spate of dark a couple of years ago
00:35:18
where right next to being able to buy
00:35:20
fairly sizeable thanks of become you
00:35:22
can get yourself minutes recharge and
00:35:24
just a couple months yeah I was in
00:35:26
Tanzania what the how to hunt together
00:35:28
is the last time together as in east
00:35:30
Africa a couple of them have cell
00:35:32
phones late raises a serious of a train
00:35:34
each fill cell phones to budget I get
00:35:36
an extra flavour a make a hundred for
00:35:38
additional cell phone minutes but one
00:35:41
thing was getting cheaper even faster
00:35:42
than computation which is an Bibles why
00:35:44
why is DNA sequencing sometime the
00:35:47
computation dropping cost by a factor
00:35:49
of a hundred DNA sequencing drafting
00:35:51
cost by almost a factor Edmonton and so
00:35:53
the question is can we I can we
00:35:55
designed essays based primarily on DNA
00:35:57
sequencing also safety but it's free
00:36:00
and that's robust enough that we can
00:36:01
deploy these these developing countries
00:36:04
to really do something about some of
00:36:06
these major problems. Um so what else I
00:36:09
would be doing these studies in human
00:36:10
eyes last runner and I'd projects maybe
00:36:13
stay and our are gets funded projects
00:36:15
our nutrition we take samples from
00:36:16
individual humans of them into my eyes
00:36:19
and then see how much weight space mice
00:36:20
gain all these depending on what this
00:36:23
and my combine the got and the results
00:36:25
we get the message you know for
00:36:26
services also supposes time this is
00:36:29
principal kind of one of the microbial
00:36:31
community with five my colonised the
00:36:33
same data used became Nassau switched
00:36:36
them why we I might same content
00:36:38
seaweed see the ship censorship to the
00:36:40
micro by you see recovery what say you
00:36:43
know database I you see if it's used in
00:36:44
the clinic but then you switch them
00:36:46
back to them why would I may cry again
00:36:48
you can model each of these periods
00:36:49
just one single exponential decay curve
00:36:53
so we should make couple paper last
00:36:55
year that we can transfer individual
00:36:56
phenotype from people in mice again
00:37:00
this is like what yes that say I case
00:37:02
of caustic already transcribe cost my
00:37:04
combine from from a kid with
00:37:07
malnutrition and mice the musty very
00:37:09
badly they lease thirty percent body
00:37:11
mass with them there's three weeks you
00:37:13
leave the country they die you can
00:37:14
basically the same you enough also the
00:37:16
supplement is clinically and you can
00:37:18
say versus couple to he change mica by
00:37:20
I however more people now suffer from
00:37:23
PC the malnutrition and so you might be
00:37:26
wondering if we can do this for a
00:37:27
beastly as well. Uh just summarises
00:37:29
briefly essentially the same type
00:37:31
transplants also work for at least the
00:37:33
so you can see there's a lean databases
00:37:35
maybe stated which is a very stable and
00:37:37
different colonisation the mice
00:37:39
functional stability these transplants
00:37:42
as a one so you can see essentially
00:37:43
paper correlations between the input in
00:37:46
the at the community in terms of their
00:37:48
PC I'm also a second marriage and we a
00:37:50
function. Um then the body composition
00:37:53
that master in a whether you have we or
00:37:55
like beast either. Um we can correlate
00:37:58
changes in particular tax or what
00:38:00
particular changes in the tabloids
00:38:02
would have let me clarify I would I
00:38:03
would really talk about that to mention
00:38:05
that we see it changes in sorting fatty
00:38:07
acid metabolism example as you might
00:38:09
expect depending on with the with the
00:38:11
tighter human this wiener index. Um and
00:38:14
then when we look across more vehemence
00:38:16
level so looking at a transcripts
00:38:18
metabolites and and number sixteen is
00:38:21
profiles are essentially what you see
00:38:23
as agreements these different levels of
00:38:25
analysis. So we think we might even be
00:38:27
able to predict the metabolite from
00:38:28
eighty nine vice versa with the
00:38:30
metadata. Um and well I was really cool
00:38:33
so we can use culture collections where
00:38:35
we take hundreds of strains we culture
00:38:37
out of single individual and
00:38:39
reconstitute Armand and whether we
00:38:41
using the un coded community what what
00:38:44
highly sorry what's all the points here
00:38:47
all the coaching community with holly
00:38:48
points after the release also balloon
00:38:50
data we get a very similar community
00:38:52
forms that individualised and then the
00:38:55
correlation on the functions is
00:38:57
extremely is extremely good will be
00:38:59
using and coach unity all code.
00:39:01
communities them but and so what's
00:39:03
fascinating about this is we can
00:39:04
finally taste ecological hypotheses
00:39:07
about what's important in getting
00:39:08
causality by leaving out the market we
00:39:10
think a really important. And then
00:39:12
bringing in the specific strains we
00:39:14
think again to have a lot to take
00:39:15
computer type So there's a character as
00:39:19
a subculture collections of
00:39:20
characterising needs to papers merrily
00:39:22
what's really cool about this is that
00:39:24
we can start design microbial
00:39:26
communities based on the lean people
00:39:27
that we can have the that we can faded
00:39:29
Jim free mouse with and protected from
00:39:31
picking up the P.'s the of these
00:39:33
people's migraines and gaining weight
00:39:35
which would normally do when I house
00:39:37
with the mouse not pleased with the use
00:39:38
micro by so we can start to get a
00:39:40
transmission of my credits good
00:39:42
correlated with these changes in
00:39:44
physiological state I I central time
00:39:48
just gonna say you this one very a very
00:39:51
quick piece of why being able to
00:39:53
integrate the data across different
00:39:54
projects is really critical say
00:39:56
associated the scrap of or infant
00:39:58
development and what's us where looks
00:40:01
like roll with smooth progression but
00:40:03
what's really cool about this is when
00:40:05
you integrate with human microphone
00:40:07
project days or so know what we to look
00:40:09
at is exactly the same data set their
00:40:11
projected into space all the data from
00:40:13
a human microphone project with them
00:40:14
out the channels in the P C.s and each
00:40:18
each each frame this one week at the
00:40:21
infant because my combine at the time.
00:40:23
And you can say that this is true men
00:40:25
just change any of my crib I'm starting
00:40:27
a but you might expect is from some of
00:40:29
the data that's the the project
00:40:30
presenters you can see some of these
00:40:32
from week to week me and micro by a
00:40:35
much larger than any of the differences
00:40:36
between two healthy adults with or
00:40:38
human micro by project despite the fact
00:40:40
we're looking at the trajectory of just
00:40:42
one person. And one thing that's really
00:40:44
cool that's coming up here just a
00:40:46
moment as to be infant get antibiotics
00:40:48
for ear infection one this is
00:40:50
tremendous regression of the market by
00:40:52
oh it in this case spiro clear recovery
00:40:56
setup by pretty quickly so just rewind
00:40:58
back so you can see it again what we
00:41:01
see what the antibiotic administration
00:41:03
is list remain just regression of the
00:41:05
market by undoing month the bow and
00:41:07
followed in this case by relatively
00:41:09
right recovery but as a problem peter
00:41:12
about double a responsible same
00:41:14
antibiotic is totally different in
00:41:15
different people. And then this one
00:41:17
case what you can see is be ones up and
00:41:20
hope you don't configure later on and
00:41:22
yet as you can from several speakers
00:41:24
yesterday with marketplace remembers
00:41:26
and Chandler intervention in these
00:41:28
really like bins despite apparent chaos
00:41:31
can have profound affect something
00:41:32
decide later on in terms of things like
00:41:34
asthma allergies and even ABC but we
00:41:38
can also do the same thing for people
00:41:40
transplants a part of the why you might
00:41:42
want to place yourself almost kind of
00:41:43
map really clear illustrations the case
00:41:45
clustered in the facilitation please
00:41:48
say P C.'s from people would see that I
00:41:50
totally different from anything we see
00:41:51
in the HMP healthy subjects this is
00:41:53
what we did without streets might best
00:41:55
be of the university of minnesota. so
00:41:57
what's gonna happen is for these
00:41:59
patients again to get people transplant
00:42:01
from the same human And so then the
00:42:03
question is how much we gonna come to
00:42:05
resemble with the data community what
00:42:07
was marked by based there. And say I
00:42:10
CDF kills forty thousand people here
00:42:12
the US align one of the most common one
00:42:14
of the most common possible infections
00:42:17
say I'm so for them again to get people
00:42:20
transpired that one day to each frame
00:42:22
in most of the day but you can see is
00:42:24
essentially immediately all the
00:42:26
microbial communities shift and the
00:42:28
healthy stays and then they stay there
00:42:31
a sentence of diarrhoea disappear
00:42:33
within the day or two at night it's
00:42:35
ninety five percent. cases and using
00:42:37
this kind of track and we can just see
00:42:39
directly affects of these my combine
00:42:41
base there is So I'm saying in response
00:42:43
to transcribe the same kind of thing is
00:42:45
very promising pragmatics spree addicts
00:42:47
and other kinds of treatments you might
00:42:49
want to apply. So the ability to track
00:42:51
the stuff them intuitive way that you
00:42:53
can explain to the patient look nation
00:42:55
I think it's gonna be really critical
00:42:56
getting forward so you are so
00:43:02
essentially what we're trying to do is
00:43:03
develop restoration ecology the got
00:43:05
which you can think of as being kind of
00:43:06
like the wall we have this initially
00:43:08
partially "'cause" system maybe it's
00:43:10
ravaged by disease or antibiotics if
00:43:12
you just leave it here so you're
00:43:13
probably going to get a lot of weeds
00:43:14
growing there seventy this is what
00:43:16
should you JCDC didn't get markets of
00:43:18
pragmatics to deeper might be great
00:43:20
good Mike recipe people addicts or
00:43:23
should you do if we call bacteria there
00:43:24
and and essentially just a chance a
00:43:28
whole ecosystem and then the question
00:43:30
is which strategies what base which
00:43:32
people but I basically to you is would
00:43:34
you get many different people there and
00:43:37
so that's really the challenge we have
00:43:39
going for what how new system inside
00:43:41
this you might combine to specify the
00:43:42
statements by individual. Um and what
00:43:46
would happen is crap and the project I
00:43:47
never got to really provide seats at
00:43:50
many different strains we can use this
00:43:52
kind of project a set a thousands of
00:43:54
members of the public been interested
00:43:55
in the sub I have to find out what we
00:43:57
get is my ideas insidious trends that
00:43:59
are not everyone's equally excited
00:44:01
what's in their anyway what that I just
00:44:05
I think whatsoever. people apparently
00:44:07
informally in my lap it contributed
00:44:09
sensible stuff any if you might mention
00:44:11
specifically during that or there's a
00:44:13
lot to me collaborations and the
00:44:15
various sources support And finally
00:44:17
thanks for your attention I'd be
00:44:18
delighted to answer a few questions at

Share this talk: 


Conference Program

Introduction of the Session 1 : The Gut Microbiome: Facts and Figures
Josef Penninger, Institute of Molecular Biotechnology, Vienna
Oct. 23, 2014 · 9:07 a.m.
648 views
The role of commensal bacteria in the gut
Willem de Vos, Wageningen University, The Neterlands
Oct. 23, 2014 · 9:31 a.m.
568 views
Q&A : The role of commensal bacteria in the gut
Willem de Vos, Wageningen University, The Neterlands
Oct. 23, 2014 · 10:29 a.m.
145 views
Gut microbial richness impacts human health
Dusko Ehrlich, INRA, Jouy-en-Josas, France
Oct. 23, 2014 · 11:07 a.m.
354 views
Q&A : Gut microbial richness impacts human health
Dusko Ehrlich, INRA, Jouy-en-Josas, France
Oct. 23, 2014 · 11:44 a.m.
Cross-talk between the mucosal immune system and environmental factors
Hiroshi Kiyono, The University of Tokyo, Japan
Oct. 23, 2014 · 11:56 a.m.
331 views
Q&A : Cross-talk between the mucosal immune system and environmental factors
Hiroshi Kiyono, The University of Tokyo, Japan
Oct. 23, 2014 · 12:31 p.m.
Introduction of the Session 2 : Host - Microbiome Interaction
Susan Suter, University of Geneva, Switzerland
Oct. 23, 2014 · 1:41 p.m.
143 views
Mechanisms of cross talk in the gut
Annick Mercenier, Nestlé Research Center, Lausanne, Switzerland
Oct. 23, 2014 · 1:55 p.m.
393 views
Q&A : Mechanisms of cross talk in the gut
Annick Mercenier, Nestlé Research Center, Lausanne, Switzerland
Oct. 23, 2014 · 2:34 p.m.
106 views
Relationship of diet to gut microbiota diversity, stability and health in older people
Paul O'Toole, University College Cork, Ireland
Oct. 23, 2014 · 3:52 p.m.
241 views
Q&A : Relationship of diet to gut microbiota diversity, stability and health in older people
Paul O'Toole, University College Cork, Ireland
Oct. 23, 2014 · 4:27 p.m.
Gut microbes and their role in malnutrition and obesity
Rob Knight, University of Colorado, Boulder, USA
Oct. 24, 2014 · 9:16 a.m.
1377 views
Q&A : Gut microbes and their role in malnutrition and obesity
Rob Knight, University of Colorado, Boulder, USA
Oct. 24, 2014 · 10:01 a.m.
The gut metagenome - your other genome
Jun Wang, BGI, Shenzhen, China
Oct. 24, 2014 · 10:19 a.m.
157 views
Q&A : The gut metagenome - your other genome
Jun Wang, BGI, Shenzhen, China
Oct. 24, 2014 · 10:53 a.m.
103 views
Fecal transplant to mine for novel probiotics
Max Nieuwdorp, Amsterdam Medical Center, The Netherlands
Oct. 24, 2014 · 11:04 a.m.
735 views
Q&A : Fecal transplant to mine for novel probiotics
Max Nieuwdorp, Amsterdam Medical Center, The Netherlands
Oct. 24, 2014 · 11:25 a.m.
Introduction of the Session 4 : Nutritional Interventions
Keiko Abe, The University of Tokyo, Japan
Oct. 24, 2014 · 12:46 p.m.
108 views
Interactions between gut microbiota, host genetics and diet
Liping Zhao, Jiao Tang University, Shanghai, China
Oct. 24, 2014 · 12:56 p.m.
465 views
Pediatric intervention - what works and what doesn't work
Hania Szajewska, The Medical University of Warsaw, Poland
Oct. 24, 2014 · 1:47 p.m.
265 views
Q&A : Pediatric intervention - what works and what doesn't work
Hania Szajewska, The Medical University of Warsaw, Poland
Oct. 24, 2014 · 2:15 p.m.
Perspectives for nutrition and the gut microbiome
Nicholas Schork, J. Craig Venter Institute, La Jolla, USA
Oct. 24, 2014 · 3:02 p.m.
1297 views
Q&A : Perspectives for nutrition and the gut microbiome
Nicholas Schork, J. Craig Venter Institute, La Jolla, USA
Oct. 24, 2014 · 3:46 p.m.

Recommended talks

Keynote: Get Ready for Change!
Michael Jackson, Global conference speaker, author & broadcaster at theothermichaeljackson.com
June 23, 2015 · 9:19 a.m.
293 views