Thank.
Ron pleasure to be here last time I read.

Georgia Tech and the State game rich.

Unfortunately and
just at last ten decades ago.

More than a decade ago also
I mean a pleasure to be.

Thanks for the invitation as I said
Can you hear me in the back OK great.

And I will kind of introduce to you in
a minute it's very different take on any

other university but
I think we do a lot of things that

are going to be a lot of interest to you
I should Lonnie you that my seminar will

consist of things that are applied
than probably goes on here although

many of you do quite a bit of
Applied research as well as.

Lot of kind of the basic details I wanted
to kind of give you more of a flavor of

different things we are doing than going
very deep into just one area because I

understand you know people in the room got
a pretty wide areas of interest here so

just how much we do kind of talk about.

A little bit and
I think that's a great idea.

To have people including
myself come from and so

some of our you know in graduate school
here or even undergrad might be interested

in looking at San Diego or any other
national lab for your future employment

you can see fairly big we have
the largest national lab.

In the country we have about
thirteen thousand people.

About you know two point eight billion
dollars and zero profit just to.

Make sure that you understand our
business model is just like yours.

And we have if you look at
kind of on the left side.

Of.
The kind of the miscellany as we call at

Sandia So it's again very applied to.

We have to be aligned with some very
specific area that the country cares about

and so the biggest one is nuclear weapons
so there is a program in the country

called stockpile stewardship for the most
of you know we don't make any new nuclear

weapons so that means we have to maintain
the ones that were built decades ago and

so that's a pretty big task and there's no
testing allowed no real testing allowed So

how do you make sure that they are in
good shape and ready available

if they need to use them hopefully never
but also at the same time nothing goes

wrong while they're not being used and
unable to.

Defense Systems and assessments think of
that Department of Defense we do a lot of

work out of op and our defense satellites
and sensors and things like that.

And as you climate an infrastructure can
pretty much guess what that would be a lot

of alternative energy research associated
with climate change and the last one is.

Pointer.

So he should have asked for that earlier.

Yeah there you go thank you.

Yeah it's on.

And I blame for
this one day international homeland and

national security where
our job is to really.

Prepare the country by doing and
be against

everything other than nuclear
weapons that we should worry about.

Everything else fits in there it's
a pretty small part of Sandia and

biology had a very broad easy to
very small part of that ten percent

roughly about a science program will be to
the tune of thirty five million dollars

per year and
that's where I fit in we have.

Kind of a big lead with him by a scientist
by energy that is actually aligned with

this group proxy here and then I was
bio defense I lead this bio defense I.

An emerging infectious disease
Pod at Sandia because he

had a lot of Engineers obviously
despite made by a non chemical

engineer because the chemical
engineering and other engineering.

And I'm not high enough
to change this by shot.

A lot lot of engineering that we hired
just to tell you that even in computer

science he doesn't have to have a computer
science background we know a lot of I know

a lot of chemical engineers who work in
that area is really driven by by the need

and not by the degree So again many if you
send there should be an attractive place

for your appointment or even
an appointment did he such a start but

he's such a staff I should also tell
you we have a pretty strong internship

program a paid internship we actually pay
pretty well if you're a graduate student

even a fancier graduate student
will get paid for summer or

full time internship at the rate of
forty five thousand dollars per year and

to whatever that will convert the hourly
rate will be the rate that you will get.

People really interested in a should look
up on this idea employment Rep A We have

a very good internship represents well.

Let's kind of down to kind
of a group that's again

in the biology we have the bio defense
I think really talked about it.

You know that we need to play a big role
here and one of the reasons national labs

play a big role in both bio defense and
by energy is that.

If you when you look from
a business perspective companies.

Should have no interest in playing in this
because there is no money to be made yet.

For these two areas can
be used later on for

lots of other purposes but
we're not there yet.

You know it's not cost effective we know
we are still doing a lot of research and

development a lot of challenges need to
be fixed before any of these alternative

energy become competent if and
that's kind of national labs enough.

Come to play in our job is to
take out those bottlenecks and

biodefense it's just by definition
there isn't much money to be made.

It's not a surprise that there was no
diagnostic F.D.A. approved diagnostics or

vaccine for Ebola because you can just
imagine how many vaccines a company

would be able to sell in parts of the
world where there is literally no money to

buy them and then again it comes and then
goes away and we don't know when the next

time it will it will happen we do know
there will be another infectious disease

outbreak we just don't know what it is so
it's a kind of a bad business model for

companies to stockpile all these different
vaccines for a lot of different things so

that's the reason why we are only here
because the kind of that's our mandate.

The Canberra national security program.

A lot of things for
a lot of different a sponsor's.

Highlighted is where I'm going to
talk about detection and diagnostics

develop them for about a fence about
toxins but it's perfectly fine for us to

move into other areas such as human health
if they have a use in that other area so

in fact and we are funded by Nestle the
sponsors who are very familiar with it and

I Department of Defense D.H.
has internal money.

So I'm going to start with a kind of.

Give you examples of two
diagnostic projects I would move

quickly through these.

Even though I wanted to bring this up
even though this doesn't fit the directly

in national security it's OK to be done
here but a lot of the technology we were

developing was very pertinent to
diagnostics and so we put a kind of a good

team together number of people
in our streets are no longer a.

Professor at Berkeley CHARLIE ROSE And
so you should be extended

at this time he was at Michigan but
he's no longer at Michigan.

You see some of the names you recognize
from Michigan and I don't regard.

Do a lot of work in microfluidics and

girl here rise to really develop
a portable device for rapid.

Think of every time you go to a dentist's
office most of the tests there are not

invasive right other than obviously
cleaning which is also fairly invasive.

Kind of good can you actually take
a little bit of saliva and while you're

waiting for your teeth to be cleaned
can the dentist run this device and

get information on a disease that
is confined to the real cavity or

a systemic disease such as cardiovascular
and you know things of that nature and

obviously that's not a ride that is very
noninvasive no one is taking your blood

and so in theory the sampling kind
of a deal in noninvasive settings.

And the second part of that of course why
you are you know everyone goes for dental

cleaning at least once or twice or at
least you should wear for dental cleaning

once or twice a year and so the idea
right that you can also get some other.

Measures of your health at the same time.

Because we don't go to a doctor most
often especially in this country.

And so the disease we chose at that
time was pretty good on her disease so

it's confined in.

Mult it's a pretty big problem
five to six billion and

this is actually old two thousand and
five data but

it's also linked to many other systemic
conditions especially hard disease and

preterm delivery and women who have
severe cases of put on disease and

I'm going to raise that right now
the top part is how we detect.

Prey down to the disease and
it's very invasive and it's mostly both

symptomatic something is already wrong and
that's why dentistry is doing a deeper

dive to figure out what's going on other
groups can we do earlier just look for

saliva and look for markers that
can catch the disease earlier and

these markers will be if you look at
kind of progressive left parietal

disease we all have bacteria in our
mouth but we don't all will develop

some of us will do a lot to do
with genetic predisposition by.

Hygiene and but
as the disease progresses you start to see

different things being released into.

Whole blood that could perhaps serve
as markers for that disease so

the blue everything in blue are kind of
the potential biomarkers and you can see

you know you know I want to put kind
of very realistic what's sufficient and

that is necessary and so you can see you
have to get to a kind of third stage at

that point kind of a confident
diagnosis can be made.

Deliver markers to now develop a kind of
a portable device that can detect it so

you look for a very simple.

You know as a red it's
just like any other.

That you probably do in
your lab to use a level and

our body you only want need one because
there is no solid fairly involved and

you have the antigen which is some other
markers that you might be looking for

if you combine the two you and if there is
any antigen in your sample you will get to

product a level and ID that has captured
the Andes and the second one which

is excess the agent and you just separate
them because they're going to be different

in their charge to mask characters tick so
you just use electrophoresis to separate

them out you should see two peaks and that
issue of those two peaks will tell you

how much of the light is present of course
you have to build a calibration curve so

that you get some quantitative idea of how
much and I did present in the sample so

I'm going to rock through that very
quickly kind of a chip that does this.

So it was a very simple chip really
feature is that we make a membrane by

photo and we make a membrane here the red
spot that's kind of the reactor that's

where most of the X. and will happen most
of the amino acid will take place and

it actually gives a lot of
advantages to using it this way so

it's a kind of homogeneous interactive but
we're using a membrane so it's a kind of.

Mix of genius and homogeneous asses and.

This is the real chip on the right side
you can see the size of the first step

is we take the entire body which is
eighty years into a little bit and

now you can see that advantage
of this memory and its memory.

So that it's very small things such as
salt and water molecule can go through but

will go through so they will stack up to
memory and literally of any little bit

about two hundred fifty or
even five fragment or sci fi none of them

will ever go through so they really
start right here in the second step and

you'll be in your sample which will be
the saliva which will have the right so

let's see whether this movie plays.

And so you can see in this case
we're just using a fluorescent

label protein to make
the movie you know obviously.

It will not be labeled from saliva and you
can see it's very nicely stacks up against

this membrane Now two things are happening
you have captured your sample exactly were

at the same place where you antibodies are
so you get very nice mixing going on and

in fact you get local concentration
is much higher than the concentration

in the same time so you actually get
an extra in fact if you do the principle

you're basically shifted it will
be in constant to the right so

you actually get much by hiding binding
constant of course both you know OK on and

off but if you do the separation quickly
enough before everything I disassociated

you actually can do Texans are very
very high level or very sensitively.

Take a step is you let it incubate for
one minute that's all we take because it's

locally concentrated there's no
limitation from diffusion really so

that can happen much much faster and
then that one minute is

really.

The field and let it go this way and
it comes up as a very nice band and

then the last one you can.

As you travel to down the street.

You really start separating So
the first really just excess die so

that we use that as an intern a standard
and then these are the two peaks.

Not separated in this movie but
this entire scale is hundred micron So

that's right you know that kind of mushed
together they're actually fairly well

separated and you do that at very
different kinds of distance of and

light and you get generated.

And against that calibration curve
you will compare your sample and so

the total time frame if you have been kind
of watching has been a few minutes I mean

even the separation gets done in about
three minutes because the scales are so

small for for capturing them for
separating them out and

that's where the advantage of time
comes in by scaling down the demands on

you really made everything much faster and

here is a kind of just kind of
quickly through that we have taken.

Kind of the peak and you know kind of
generated a calibration curve with non

standard and then we actually put some
patient samples here you know the patient

from pretty down to disease came from
University of Michigan dental school and

you can see that I'm
right on this curve and

we can figure out you know what
the content is of this particular.

Market which is M.P.
eight it's one of the and.

The concentration goes up if you
start having bigger it isn't and.

You know obviously you know
probably a lot of papers.

And we had an integrated device with this.

But there was also a lesson learned and
I was going to share that

since I talk about Applied Research So
this is all great for publication and

things like that so we then approached
a lot of companies hey this is so

cool you must be interested
never just learned that.

One page for Peter insurance companies
don't pay unless it's really bad and

they have so hence they have no imperative
to detect it early because then

it's really bad news for them they
have to start paying even earlier.

And so they just ignored it and
so that's number one so

don't start developing technology unless
you know that there is a need for it.

And the second I had read this
is much interesting Rand.

Of the doctors are not happy with dentist.

Doing diagnosis of systemic disease
because that's their turf and

I know I'm being recorded.

But you know but it's a very important
lesson you know maybe it's not fair and

maybe to Shouldn't be that way but
it is that way and

you have to be really careful.

The dentist would love to detect
every disease you ever have

right cardiovascular and things like that
but then doctors are not very happy so

now you go to a clinical trial it becomes
very very complicated as to who you

partner and happening with both of them
at the same time is not that practical.

And because both of them are individually
happy to work with me but not so happy to

work with each other and these are kind
of the lesson learned the alternatives

to the enemies that I forget put on
disease there's a lot of other things that

should be useful to other fields as well
because if you can do put it on television

marker like M. and B. eight is just like
any other protein the problem of this is

a little four document absolutely
great in the lab and great in running.

Samples that are fairly clean not so
good for doing serum and blood

especially not so good for saliva either
because it could be a lot of various sums

in terms of you know what the protein
county's what the total blood count is and

that actually ends of shifting your peaks
and so it wasn't a very practical devise.

That we learned when we started
applying this to other applications so

even though it was very exciting we were
very excited and didn't go anywhere and so

that's kind of you know on
the breast twenty six or

seven years of my you know heart of my
career working on it Lesson learned and

lesson learned that it's never
too early to think about.

What the technology can do if that is
hundred percent successful problem solving

and I tried to teach this earlier to
people who come to us and the unfortunate

Unfortunately there wasn't anyone
else telling me when I joined so.

Let's move on to a spin direction you
can see how simpler this device is

this is a lab on a desk
again not a platform that

we invented centrifugal
microfluidics has been around for

quite some time except that this was
invented because we had a problem.

This made perfect sense to develop and
get to you in a minute as to why

we have lots of patents and
just to give you the news.

For the last light had been commercialized
by six companies including one a startup

from a couple of staff members on
a postcard that reported to me

left to start their own company and
they are going after neonatal disease and

other other diseases and in addition
to that company there are other five

companies that have
licensed this technology.

So I want to acknowledge most of the work
actually was done by people listed here.

SCHAFFNER post-doc and great summer.

Staff member who left to
start their own company and

the concept of it is very simple and
shown here in the bottom correct

transfer to a disk or
disk into this device and

then run it we wanted it to be a plug and
play you know centrifugal microfluidics it

sounds very much like CD player actually
that's where the idea came to people.

And number of people have
been pioneers in this area.

All the chemistry is in the disk so
it's just like a CD player the magic is in

the days player is just a passive reader
which will read any particular disk and

you have to basically make sure
that the disc is encoded so

that the player knows what it's reading.

You know you really need to add to
twenty Michael lead of samples and then.

The reason is that you can do her blood
directly into this device that is

no reason for the blood collection you
could do that if you have access to that

that can be done as well but
our garage takes a life or

her blood directly by either spitting for
saliva or by pin prick for

her blood put it directly in the reader
and that should be enough to run this disc

and you know the time and things like that
the disk a single use obviously you know

you you know that's where it's unlikely
that you can use multiple times but

you can do up as many as
sixty four at the same time.

You know based on amino acids just like
the last one except we use beads here so

you have beads.

That have antibody attached
to them that's your ears and

we also have a little bit about it so
that's your secondary agent.

Typically quantum dot because they're much
more stable and we add them together and

here is recreate the novelty is really
just going to center field microfluidics

by expanding the discrete actually
prepared density media so

this is like analytical center figures and
for those of you where you

have done certificates and at larger
scale so and then we pick the beads so

that they are the only things heavy enough
to go through the density media everything

else everything which is very close
to one point in a specific centrally

stage on top you know because we can put
we can make it close to one point two five

nothing in your blood or
any sample will go through this top

to have done my sample prep while you're
doing that you know asses beachside silica

to point to much much heavier specific
gravity they're going through but

by the time the growth through they
have actually captured the and the light

as well as the label antibodies so you're
kind of formed a sandwich which is for

us and and and but the only build and
her body again it's a protein

specific gravity very close
to water straight on top.

So you have done your sample prep and you
are separate sort of bound from on bound

in one step and then you have
multiple antibodies so there you go

again another constant is an effect that's
going on each beat captures many many and

like molecules at the same time and then
you have thousands of beats so you don't

have to worry about sort of leaching and
things like that you can read this for

a second or two seconds and
don't have to worry about for

the bleaching because you have plenty
of concentration here to read.

It really well again I wanted to
point out that there are many.

Devices out there the Center for Grayson
out is very different because none of them

really use that density media
that we put in and that's.

The patents and
many of those patents have been issued so

companies are happy to
license this from us.

That's another important thing
you can't just do it for

the first time you have to let everyone
know that you have done it for

the first time by falling apart and
probably get some doesn't really count

when it comes to a commercialization and
so we have.

Got to do that very aggressive about
protecting our IP because if you don't IP.

Minimal chance of anything
being commercialized

because that's what going to protect
the company from using your technology and

protect it from anyone else copying
it here is an example we were talking

to I'm going back to my
bio defense kind of.

That.

It's a big problem.

You know you get white powder samples
you're probably already of that how do you

know what toxin is in there it's the most.

You know it's the most pox and some
substance known to man I mean literally

a few nanograms can kill you and so you
detect insensitivity has to be very very

high to detect it at eleven concentrations
and here I show you that we're detecting

it and I'll show you the sensitivity in
the next slide but you can see we can mix

of course talks and is not found
in all these food in a person.

Kind of you know became very ambitious and
basically took everything that you could

spike them with toxin and basically sure
that we can detect it in pretty much

anything you can think of obviously
it's very tough to get canned meat and

peanut butter so there was some of.

That we had to do before.

The second part I also want to tell you
this is another important thing that

you are developing for the first time or

whatever method you always have to show
the community especially the biological or

clinical community that it actually
matches the threat being used now

that's the only way you can make them
confident that this actually works so

we actually brought commercial
actually is not a commercial.

Developed by U.S.D.A. because they
have the mandate to test food for

contamination with botulinum toxin and
blind and you know when we do very very

well compared to U.S.D.A. approved.

Here to when I was going to show you for
the sensitivity so here is the grill

a standard for botulinum toxin on
the right side beyond ten people and

it's a mile marker staffed mostly by asset
so literally you take your sample and

you feed it to these poor mice and
you see what happens to them and

they have to sacrifice them no
matter what happens to them so

even if the toxin you have to sacrifice
them at the end of the day and

here is kind of the result you get You see
we don't run that many concentrations for

mice because you don't want
to kill that many mice.

You know you're a you're a limit of
detection is somewhere between two to ten

although we couldn't detect any
ten sometimes you could it also

depends on the mice there
is another one called.

West so literally if you give them toxin
to my eyes since it binds to neuron.

Turn into a dumbbell.

It's not a good picture to visualize but

that's what happens to them the wrist area
becomes really ten and that is very is.

Pacific for a bottle and I'm talking to
people can use that as a subjective test

as well but needless to say we can we do
much much better because we are really

even the last concentrates on.

It.

You know like hundred times
lower effectively even if you

take over ten factor.

We are ten times better
than a mouse by Lassie so

that hundred years are is much much
cheaper and second you are not sacrificing

my so we are really hoping that I said
like this can replace the most testing for

many of the toxins that get gets done.

It's not limited to bad defense
toxins only how mean we can do.

Sixty five in because if it's any
meaning we also want to see whether

this is a genetic amino acid platform and
it really is.

You know again I only talk about
bad this was actually published in.

Chemistry just recently we can do
multiple toxins at the same time and

so here is kind of an example where we
have it around Rice in protective and

isn't she.

On the on the same disk and we do and I'm
going to be compared with whenever we get

a chance and we can either meet
the sensitivity that is offered by the.

Times of the actually do better.

So here's a question since there
are lots of students here.

To scale the wrong to go wrong and so
here is an example showing you that.

You get a kind of like this and so
maybe at the end of the seminar

if anyone has an insight as to
why this happened let me know so

he what's going on is I have to
concentrate on just think of.

And here is your X. axis which has an
increasing amount of you go from left to

right and left is that it's coming from
binding and you get a kind of like this so

I would like to know why do you get a kind
of like that obviously is the problem

because if you get a kind of like this and
if you look at France you don't know.

The contents only see one or two and
there's a huge difference because

he too might kill you when might we
just OK so anyway this is a kind of

small Christian if you don't have answers
that's OK I'll tell you anyway and

this is real and
I will tell you right this happens.

We can answer since it's a simple
after you don't have to be limited to

you know assets any place where there is
affinity binding you can use it and so

we shorted for.

For Food borne or water borne illnesses
where you don't have to detect very.

Little concentration and
if you especially if you concentrate them.

Actually works quite well.

I think I'm going to skip through
this fairly quickly to kind of

get to the last part.

I already showed you this it's compatible
with blood in fact actually you can see

the red blood cells I don't know whether
that's visible to you it's pretty

nice on my screen actually you
can see it spect right on top of

the density media so your center figures
and you can actually use it for sorting or

not but self-righteousness and
if you would like to.

And he's going to tell the difference
between red blood cells and

white blood cells by the way and if you
use anti-bodies you can subtract an eight

right blood cells and
it's a very simple platform

very easy to adopt in your
labs if you're interested.

Very simple two pieces of
plastic stuck together by.

Double sided tape.

And so we take a little cut the tape and
that defines your channels so

all the channels you see
basically you know so you can pick

any thickness of the paper you want to
that gives you the depth of your disk and

then how right you cut it defines how much
total volume you'll have in there and

then literally read.

On the top part of.

Just with any C.N.C. machine and then we
stick them together and you have your.

Of course if you are scaling it up you
most likely go to injection molding or

some other other method which is
more scalable right I mean this

is something that person to do.

Or do you see it and
then stick them together but

even in our hands we can make
hundreds of these in a day.

And a little cutter you can buy for
about fifteen thousand dollars and

then also make them someone just told
me that they made it by three D.

printing so that's a possibility
too not an expert in three D.

printing so
I don't know what they used a very simple

disk is less than two dollars even with
the anti-bodies the cost is so that for

honesty the cost is not determined by the
plastic because of course plastic is cheap

cheap cost is going to be determined
by the cost of fuel use and

which of the anti-bodies and labeling them
how much labor is involved devise is very

very simple like it's much
simpler than a CD player so

in theory it should be less than
fifty dollars However you know.

Actually a person of the CD player they
are only interested in making something

like this if I order one hundred thousand
which obviously is not the case so

in our hands it's cost about
two thousand dollars to make it

that's just the cost of
the material if you include labor.

It's much much higher.

OK so if the question kind
of move to the last one is.

Kind of my preferred IT platform for
synthetic biology probably different.

Bio Energy and I will introduce that.

In a minute.

Synthetic biology is
a fancy term it's just.

Like engineering or
genetic engineering by other name.

And here what I'm thinking about
is putting entire Patrese.

And moving it into equal
least at one time.

And kind of get to that sort of.

In a microphone platform most of the
record has been by a number of stocks and

research a staffer J.B. Steve she and Phil
actually have done most of the work and

then Peter and others have helped as
needed just is no longer our job left for

a start up company.

I think I don't know how many of you
are familiar with Byron A Z There are only

a couple of ways to do bio
energy the big areas are.

So kind of biomass you know
Conestoga are not corn

in the world include corner here and
the second is.

That they're starting and
this is more the plant like trees and

grass and things like that.

And then you have multiple steps involved
you have to first break down the biomass

kind of pretreatment very chemical
engineering typically do it combined

chemical or heat and genetic then to
release the senators and then you have

to bring in because that's the only way we
know how to do this very efficiently that

break down those cellulose to simple
sugar which then can be used by microbes.

To make biofuels Pamela's here obviously.

And I were together have question about
synthetic biology addressed to her

because you know that scene or
that much more than I do.

And again kind of introduction to that
might be also of interest to you it's.

Kind of introducing that in the beginning
it's a multilevel partnership so

this is funded by do you eat and
do you fund I think it might be the only

agency that fund things this big
anymore especially given given and

I just lack of interest in
front of the example of.

A little limb is twenty five million
dollars per year we got another

thirty million to buy
equipment right at the big.

And funded for five years at a time and
we are in the second renewal right now and

we're hoping that we'll get another
third in two years from now it has

seven partners for
National University and one foundation

kind of there are listed here
Lawrence Berkeley is the lead lab.

Another chemical engineer from U.C.
Berkeley Lawrence Berkeley

he's the lead Sandi I was part of it.

For a lot of plant biology and
we have Carnegie science from

kind of the Stanford campus and
Pacific Northwest National Lab.

For been isms feedstocks deconstructs and
fuels and cross-country technologies so

I cross-cutting technologies out where
our goal is to develop technology

to meet the needs of
the other three divisions.

And.

That kind of brings me to kind
of ride microfluidics an idea.

You know really regular fairly
applied project screening becomes

very very important because even
few percent optimizing compared

to existing method can mean a lot
in terms of your total cost

on a scale that you are shooting for
and one way to do that

is to screen lots of different conditions
just to find that best condition.

That the you need to have and that can
be for optimizing your enzyme cocktails.

Even a concentration difference
of ten percent can make and

I'm talking ten percent cheaper or

more depending on exactly what
particular protein you're talking about.

Kind of this Internet biology area where
we are trying to put cloned bunch of

pathways.

Back to a yeast or a fungus and let me
just give you an example of this scale so

suppose you have a partly of six genes

that you wanted to put in equal
to make beautiful or ethanol.

Not that interested in ethanol anymore but
it's a beauty.

And suppose that Teller two hundred
species very easy to get you can just

mutate a couple of time into acids and
can get to ten alternatives or

they can come from different organisms
because many different organisms have very

similar enzymes.

So if you want to do an entire communism
you have to do tend to the six

experiments so
that's a million experiment.

Almost impossible to do in a micro
tighter plate because that is our

standard high throughput
platform available.

And just think how many micro Tiger
played you'll have to run it's almost

impossible to do it manually of
course you can buy it really bad

by the now you're talking about millions
of dollars to buy the robot and then

multiple people just babies that are about
because they don't work as advertised

that's all been there has kind of you
know and we haven't gotten there yet

can we should everything down to take
advantage of it scaling down to do larger

number number of reactions
in a smaller footprint and.

Hopefully also make it faster and the most
important by automating it you take out

the human error from that and
the last one is that if you scale down.

There is significant cost saving and

that's actually the bottleneck right now
people can do a million experiments for

synthetic biology part
to synthesize D.N.A. for

those million million experiments would
be cost prohibitive we would just not

do it because it could be in the orders of
hundreds of thousands of dollars and so

the cost of the biggest driver you know
bring it down by a factor of hundred and

I do write that just use this droplets
these are water droplets suspended in

a while as discrete reacts and
chamber just like your micro tractor Well

if separated by plastic he
would have separated by oil.

We were not again not the first
one to think of this so

here are some groups that have
done very nice work in this area.

I think you started it
in your city of Chicago.

David.

Came from White's lab.

And there are many others just just
to give you an example of this.

Attention from a lot of people this is
a format where droplets always keep moving

so we kind of call it droplets in
channelled operating flow or analog.

Microfluidics and I know.

That using something like this we also
used it in the beginning to do lots of

friends I'm screening at the same time and
so here is a paper from lab chip where we

actually use solid biomass to basically
use the Doppler to do her genius.

Which are very hard to do in a micro
plate because it's very hard to dispense

solid biomass most of you are chemical in
the newest you can you can easily figure

it out that is very hard to dispense
solid biomass we figured out how to do it

in droplets and you know problems however.

There is a letter from that.

Droplet microfluidics which is
something called Digital microfluidics

where we use electro wrapping.

I don't know exactly
what the mechanism is but

by activating these electrodes you can
really concur the hydrophobic city

of your surface and hence also you're
changing the contact angle real time and

make the droplets northbound and
you can see you can you can type it so

you can see this one is being prepared
to start with the big droplets

you can just a small property equivalent
of by padding from it as it arrives and

then you can miss them and you can see
that movie being generated there is

actually a company called logic
came out of Duke University.

The big sequencing company recently so
they're doing a lot of this work so

don't think that we have the first one to
do this you know we're just using it for

different application and
the advantage of D.M.F. is that you have

absolute control over things like
mixing making the droplet sartin.

The disadvantage is throughput
because you can only do.

As many unit operations the number of
electrodes like are fewer than that and.

I think if you want to do ten thousand
reactions you have to have a minimum of

ten thousand electrodes that
you have very expensive.

Going to college that's not expensive but
here for

microfluidics that's very expensive and
we don't really know how to scale that up.

That ladder for droplets in China rise to
speed and so our concepts can be combined

the two so take the high throughput
advantage from droplets inflow and

let them keep them flowing and we don't
need to do anything fancy to them such as

give me a sense things like that because
they don't need to sit on an electrode to

culture but when you want to mix the two
things like you know added about it

to your culture media and
then we bring them to an electrode and

we do that sequentially so
I want to you know one drop at a time and

then put them back in the channel
again to incubate for however long

long the need to include it and has read
quite well is still a long way to go.

And these are the kind of unit operations
are molecular biology operations we wanted

to do will be just where everyone from the
beginning start with the D.N.A. part so

you have that example I
showed you spotted of six.

And ten Brussels of each gene
you still limited sixty so

having sixty inputs to a microchip is not
that complicated if the Communist come

in a totally space that gets very large
because it becomes ten to the six and so

we wanted to do and

was mixing of the other parts together
with the plaza made on the chip.

And then do some culture because at
the end of the day that's what you get

out of the chip is the cultured cells
then obviously it has to be a scaled up

we want to test.

The number of different conditions we
want to make sure that it works with many

different techniques that are commonly
used by people doing metabolic and.

Biology such as Golden Gate Gibson
you know don't ask me why those names

are you know they just happen to be.

People who use them because one better
than other depending on the application

but rather to see that
all of them can work.

Under Dick this entire box and put it
in a chip or multiple chips for now so

here's an example just published in
the synthetic biology where you can see

you can also see the channels so it's kind
of the combination of the two formats of

microfluidics And so we do all the D.M.'s
only because that's all mixing that's all

merging of droplets we do them all on
electrodes but when that had been done

we let them do in this kind of challenge
and that's where they have the.

Operation going on and
you can make a second dying to just

add the residence time without
making the chip bigger.

And then we have another set of
electrodes to bring in the cells so

we can do the electrical transformation
and then we have an outlet

which we also have some built in volves
to make sure that there's not going

out when we are not ready to but we can
also extract cells out of this chip and

do a kind of regular culture and
this just shows you.

That we can actually achieve all those
kind of medical molecular biology steps

it actually works pretty well we have
shown it found all these different methods

to get that has happened at room
temperature Gibson fifty to get a better

as we have a peltier heater at the bottom
of the chip because obviously if you want

to do it at fifty degrees you need to
have some kind of heating element but

you can just take literally
bringing a peltier unit

to the backside of the chip and
that's good enough and

then we did some modeling just to make
sure that the heating was done right.

And we also did used we did.

And we also did.

I think fungus the pipe and

again through these quickly you
can see how different a step.

You know I read I did
just to make the movies.

And kind of that I got them shows
you that all you have to do is

have a good control software for
very dangerous for the electrodes and so

this part better get complicated Right I
mean if you have that many electrodes

you have to now run lines to.

Get them out of the chip to connect
to a pain or something outside so

that part that's got complicated and
kind of has a limitation and

scale up in a sharpie we
just did kind of sixteen.

Droplets and
sixteen different combinations very good

sequencing match for
the I guess on reason.

And in the last part of the chip rather
transformation not really we can do

a little present to
the last chip we did and

I like to present I think I skipped it
too quickly so that we kind of format

a bit of the curves fairly close to each
other so we can generate high enough field

to electroplate the cells we can also do.

Because that's very standard we wanted to
make sure that can we develop a form of

a chip where we can do huge shock and here
is a kind of chip on the right you can see

the different temperature regions
forty to thirty seven zero degrees so

you can kind of a stirred things up to
whatever the right compressor is for

each of those just steps
are gay rights right where.

We can do you know multiple
different type of protein again.

Not to be but there nicely for us and you
can use that as an example the same chip

can do all three we did hundred Doppler
and they look almost identical so

that means there is not much
variation from Doppler to droplet.

And we also did equal I used a fungus and
you know it was

compatible with all three and I'm certain
I'm not that that impossible is just you

have to test your chip with many different
things that the biologist would like

to do to make sure that it actually works
not just with one type of cell which is.

Because whether someone is
interested in East or Niger and yet

they were actually quite well in this
keeping a lot of details we had to

do a lot of stuff as chemistry and.

Anyone who does droplet knows all these
kind of pain points and so we didn't spend

a lot of time optimizing a lot of those
conditions before we got to that.

So where are we going next I want to
kind of leave enough time for questions.

If you.

Don't even have to notice they're
all kind of different we did one for

here chalk one sort of you know.

To actually finally integrate everything
on one chip all you do is have

these sixteen lots of sixty.

You introduce Not sixty three million
sixty because you have to have like

rushing before and things of that nature.

But then everything else gets done a
device that looks something like this and

has tons of Rives obviously coming in and
some outlets for the liquid and.

Probably liquids out you can also
notice that this is another thing we

pay a lot more attention
at a national lab is that

this is fairly well engineered
in terms of we're not using.

Stock in a pretty as Chip as our use kind
of realistic involved in things like

that so that it can be eventually be sent
to other places places other than J.B.

or if a company is interested you
know they have actually a decent

a starting point in terms
of engineering itself.

And I would usually just kind of again we
need to be a little bit more applied than

he was that is so we tried to bring it
closer to where a company will pick it up

and that the first challenge.

You know and Pam is working on this
to you know we need some really good

idea there to actually to be a challenge
as well that is listed here philippic acid

How do you know that your cells are making
what you need them to make because

most of them are not going to be for us
and it's easy for me to show you P. and B.

if B. and B. and all that by.

No one cares about those things.

So if you're making butanol
especially on this a scale

it's almost impossible to detect.

Droplets from Nice ideas about
can you have a proxy for

that compound being made.

Investigating connecting the Doppler chip
to him I suspect very sophisticated ma

suspect platform developed by
a colleague of mine a lot of Berkeley.

But I think there is room.

For improvement and
second part of the biggest challenge is

establishing a scaling laws just because
you can do the communism here and

it works really well how do you know that
that would be true want to scale it up

because eventually whatever it is that I'm
making in my chip needs to be a scaled up.

If not more to make something real and
if it's bought by a fuel

that will be you know we're not
thousands and thousands of gallons.

So how do we know that just because
it works on the chip will work or

just because it fails in the chip it
doesn't work at higher scale and that's

another thing our grill actually is not
to do it in the chip we do it in the chip

than we do it in the micro tighter
plate than we do it in millimeter scale

then need to scale flask to also started
stablish ing scaling large as we are kind

of shrinking down the assets to see you
know what those labs are that govern.

Taking things from this a scale to
have hundreds of liters of scale.

Stop their.

Tank most of these people anyway but

obviously it took a lot of money from many
different sponsors to make it happen and

I listed some of the grounds that
contributed to this to thank you all and

I'm happy to take questions.

I.

Yes.

Very good point but was that two slightly
different aspects what I meant by this

is that as you do that everything works
perfectly in your droplet microfluidics.

How do you learn.

That once you take it
out of the droplet and

your knowledge of culture between the same
right and that this is one of the big

microfluidics is that if you want to
scale it up you have to establish that

things work that's I mean you know I
showed you in a meeting last year and

I'm going to be doing the chip we compare
it to rock the greatest and that is now so

you have a similar level of effort and
bring in some good model.

Because if it if everything works
like a charm and everything I do and

operate it works just fine at hundred
leaders scale that is nothing to be solved

but we all know the answer that
it would not be the case right so

then you need to figure out what is the
problem arises not a scaling up because

it's not you know the molecular
biology should happen the same way but

we learned that in a culture in a.

Test tube is different than the culture in
the micro Dr Well we should different in

a culture in a hundred or floss and can
actually i'm lot of that it has to do with

chemical engineering right I mean it's
the lots of other things that come in and

that's what that's what I meant by
a scaling laws rather that that I'm

not reproducing rather left in my chip so

that my results are very different than
what you will get at a bigger scale

the second baby that's a very
liberal because yes you know.

And that's also not solve the yet
I mean there are a number of companies

that are trying to do that but
it's all very early stages and

they're all focusing on the market.

Reactions and be the one
millionth reaction identical to.

The first reaction I had and that's a very
good question because by the time even if

each run even if you do have a frequency
of can hurt there is a huge time gap

between the well the first droplet and
the millions how do you control for

that right I mean you know
how do you detect all the so

that you have so
that's a lot better scaling problem but

very far from this we just want to show
the proof of principle around two thousand

I think companies that I want to come in
and then take that second part of your

question.

Yeah.

Yeah yeah.

That's pretty easy because typically
you are going to do what you can

do you have to interface with something
that will take you Michael leader so

in the studio I showed you can
type in directly into that

that's not a problem when the center
think you have a very good point so

just to give you an example of
how companies are doing it now.

If you look at.

Device that is being sold by Illumina but

there have been a typically they have a
micro tighter plate on top of their chip.

So it looks like just like I got
up late and you go directly so

they have this plastic manufactured that
is cheap enough that they can put it

right on top of the chip and inject right
there because that's much much simpler

than coming up with an interface to have
something like a micro hydroplane it

can connect directly to the chip
that's much harder because of the.

The volume in compatibility
you typically by putting in.

Your electrodes when you need sub
micro leader how do you do that

very accurately so that's still a problem
some kind of answering it in a long way

that's where one of my kind of biggest
issue with how this field works

that's when we need to
develop some standards.

On the community we need
to figure it out and

run it will be how do we interface with
micro tighter plates they're not going

away they're used they're very useful and
there's a lot of money invested in it and

it's going to stay there so I would say
for the synthetic biology chip dancer will

be a make I need to make my
device compatible with a micro.

Because everyone knows that
everyone can give you D.N.A.

pieces on a micro plate and then and
both input and output right

because if I gave Pam an output which
had my provider played with cells

she knows exactly what to do with that
next and so I think we need some extra.

You know the other option
would be I make the inlet and

outlet customized on the chip and
that will work but

then you have to have enough market that
you can you know because that adds cost so

then you have to have can you really.

Have that cost in the price
of the disposable for

sequencing it's OK you can do that but for
a lot of other applications I'm not so

sure so they're not really
depends on the application but

I think you know in some way to
be nice if everyone just agreed

that their chip will work with
A three eighty four plate.

Because you know that's a standard and
everyone has it and

I think that's something we
have not solved as a community.

OK.

So.

The mission but it will be and probably
professors would not like it ask hard

questions as to why you are working on
the project if it's a rock solid thing.

And you know I will tell you my Ph D.

I mean I just rot in my professors and
this is about doing the work on and

I was just so happy that I had a project
which was founded I didn't ask any

single question I did because there was
another that everybody outside my advisor.

Who is happy to jump on that project and
so maybe you don't after.

Getting assigned get assigned and then
professor knows that you know that I stuck

with you and then you asked that question
right but it's a very little and

I mean everything is available if your
child walks out of it going to go

right I think that's something I
learned very late in my career.

Second when it would be.

You know just kind of a scale
of things how I live the very

best expert in using
this device in my lab.

Well once you leave can anyone else use
it in the same lab or somewhere else and

to put you know put some
attention to that as to.

How practical it is and if it's taking you
three days just to get the chip up and

running after the tenth I mean of course
in the beginning it takes three months to

get the ship up and running but after
you have the power lines you know from

the perspective of who the end user will
be so that two of them are very connected.

And kind of goes back to you in
that particular question right so

we think about that very seriously now and
one thing I've realized there is no good

going to solve the problem for me but
I think that would be pretty pretty useful

to be used modeling I mean that's all the
third thing I would say modeling saves you

a lot of time if you can use it and
there's some really nice packages

that are already out there which have been
very very user useful for modeling and

simulation especially if
you're doing kind of micro for

that kind of thing they're going to teach
you a lot more than you think and it's

a very go to scale that it fairly widely
needed like national labs and companies.

Most of the national labs have.

Licenses for
the spend most of the micro for

the packages because you can afford it at
the lab of thirteen thousand people and so

if you want to come use it it's almost
free for you and really look for

that skill set because it's impossible to
hire a model or just for your project.

And if you're good at modeling then and
stimulus and

you also have a better to medical
understanding of what's really going on.

Thank you.