It was a very probably calculus just
a professor in the department of chemistry

and biochemistry here in Georgia Tech
come to receive his Ph D.

in the chemistry for pointers are a
University of British are a series of

American imports it gets better.

I could be subject to
the postdoctoral researcher for

short time Stanford of the Laboratory
of the exhibit was world class for

so he's got an excellent pedigree and
then he spent some time in

Minnesota actually in the University
of Arizona during Georgia Tech.

He had a separate board a plane trip for
Georgia Tech.

He has American Society for mass
spectrometry research award if you want to

reach forty first class publications three
happens in an excellent trajectory for

the most important thing about from up
from that was like urge you to Google

his name modified hottie
professor that I know of

you have such a distinction in the mission
for a very impressive Electoral right.

So I'm not going to ask if this
is how supplements the grid

is just new wrapping molecules.

But anyway.

Today it's a real
privilege to introduce for

the one that enabling me
a spectrum of technologies for

your mix of our Thanks Jeff.

So carefully look at my funding
line at the very end and

you may find some I'm going to
be a Web site talking about.

And that's the first thing the Pops How do
people in my name and it's not my phone.

But anyways if it's something from
stand for that has a lot of free time.

And so it's a real pleasure to be here and
see enough friends here in the audience

and it's a real challenge to show you
what we're doing in any light that.

Appreciate.

And as a see me if you know my group or
my students.

I feel very proud of working with them.

It's there it's a very dynamic group where
a small group they get a lot of things

done and that's the way I like to
define it and we have more ideas

than ours in the day and we collaborate
with a lot of talented people on campus.

So I'm going to going to
go through the start.

I'm going to show you what we do with
different people there are groups on

campus and rather than giving
you a pretty and nice picture.

I wanted to show you the challenges.

Indeed these fields and.

I guess you're tired of seeing
talks that you know feet.

So well and there are no challenges
no unanswered questions is not

going to be the case in these times.

So if you're sitting in the silence I
assume that you have some sort of interest

for the world systems and that that
could be system set approach scientific

research as a general thing or it could be
some sort of systems biology environmental

systems biomedical systems but
what's common to all these points from

our perspective and a little cannister
must be trying to trace is that all these

systems disciplines rely on something
that I like to call system analytics

which is the ability of look at the system
to Michoud the system seem time and

space and how they evolve
over certain experiments and.

Certain mentally ticks it's a disciplined
I likely find and because conditions for

many other subdisciplines for example
genomics I would say it's the first

systems and only takes field if they're
throughput feel that it's really

dedicated they're looking at genomes in
large scales in very large numbers and

many many gifted people in this
campus are working on a general mix.

To maybe your fields are proteomic summit
our law makes projects I wouldn't say so

you know it's a fairly mature feel now.

Depending on how to define
where the proteomic started but

I would say that broadly it
exploded in the last ten years.

And so at the same with a low mix and

middle of it is not so
easily find these days and it's.

It's the best word that you will see
coming up in most conferences now so

people have moved from genomics to
proteomics to meet our longings and

who knows what the Next on NEXT will be
but the common thread of all these topics

is that we're looking at systems and
we're looking at making measurements

accurate measurements of these systems
in very high throughput manner.

So just because they love to the fine
things before we jump into and

what is proteomics Well M.
the classic view of proteomics it's

the that it includes the an indication and
quantification of proteins but

it's not only that that's just what I like
to call called brainless proteomics OK So

just like the classic crunch sample
after some pull proteomics but

it also involves determining the
localisation of proteins how their money

file how they interact with the final
idea of determining their function and

I think that picture is much better than
any finish and so I always loved this

picture in and there are many differences
between the Department of genetics and

biochemistry and the Cylon department
of genomics and proteomics and decided.

Of course this is not a hippie ish
scientist longer hair looks cool like

sitting on but the main difference though
it's the type of tools that these two

studies start using reinsert with
a cane we see very nice activities but

it's going to fish one
protein at the time.

The scientists here is
fishing with a crane.

So it's going to fish a bunch of protein
study time and that's really the change

between proteomics and protein
biochemistry that doesn't mean that

this approach is not useful and we have to
rely on these techniques at some point.

So these are really
complimentary approaches and

proteomics the subtle enough.

It's an acronym reach field.

So if you've ever tried to get into a
field of preventing So you've been reading

papers for the first time that you
face is this type of things workflows.

Acronyms am and there are many
ways of doing proteomics but

the simplest way of identifying
of identifying a propane

is by digesting that protein and then
we must pick from a tree on the protein

fragments identifying the masses and
then doing a database search.

So from its initial times.

Proteomics is a very disciplined
that depends heavily on a chemical

instrumentation and be computational
power and this too can read together so

that to be unique or odd proteomics the
techniques that are used are not new for

many a few techniques like
Molly like Electra spray

little face suppression techniques
legal chromatography and so on and

many different search algorithms say it's
very hard to keep track of their infinite

number of soft protein search algorithms
are out there but these were close have

been shown to be really successful in
the N.T. fine and comparing the the.

Quantity of proteins in
different systems and

from it alone makes I have
a hard time with this line.

It's hard to find and
I agree on all definition of meat.

Our long experience and
even agree on the name of the field.

Some people calling it our law makes
capital going it up or no means am I would

use the word Metabo no makes when we're
doing quantitative experiments but

the basic idea is that in a bit
meet our long economy there

were no me to explain why looking at
their comprehensive on quantitative

analysis of all neat double lights
in a system in a given system and

that could be done to understand
the system to understand how the system

changes to investigate new drugs or
to for example do that.

No C C And that's one of
the things that we are doing.

I'm going to show you that in a second.

What it really boils down to is that
if we want to do genomics proteomics

from a double nomics right we're going
to need some sort of new tools and

that feedback between new tools and
new fields.

It's actually very hard.

It's not that new questions generate
returns but some some new tools in able to

answer questions that we didn't you
think they were possible before so

this is really a two way
relationship between the field

questions hundred tools and the tools are
not only experimental like the ones I'm

going to show you are also computational
So it's really a two way relationship.

And this is just I promise
I have no author rights or

licensing rights on this book but
it's just a book that I really like and

I some of you may have
some time to read it.

Maybe take my.

That to work you could try to read
this book it's called Thing knowledge.

It's a ten Interesting theory it's
a modern philosopher who basically

says that.

Scientific Instruments have knowledge and

instantly in themselves
they contain knowledge.

It's not the knowledge they generate but
it's a way of transmitting knowledge and

it's sometimes regarded as a lower form of
knowledge but acknowledging themselves and

it's a really interesting and
he gives a few examples but for example

I think we all agree that the human genome
we know you can have sequenced the.

Without a D.N.A. sequence are OK and

nanotechnology will be passed on without
powerful microscopes and so on so that

that really points out that that synergies
and between instrumentation methodology

tools and scientific questions coming from
all the old mixed fields and so what's our

scientific machinery what I want to write
about surely works on developing machinery

or developing tools and methods but what
are those methods as you know already.

And there's one more that we
can get rid off and that's

mass spectrometry we do a lot of work with
Master trauma tree and must become a tree.

It's basically a simple way of
measuring it's a molecular scale

it's a way of measuring the math
situation of molecules and

the way you do that easily manipulate
molecules with electrical and

magnetic fields and you can move them
around and you commission arrival times

you can make sure frequencies of motion
but the bottom line is that you want

to measure in Massachusetts ratio of
those molecules with very high accuracy.

Maybe with four or
five different places in.

But that's only one component of
a picture and I wish it was so

simple but in order to do must be
a trauma to First you need aisles and

aisles are basically gas space molecules
would have to have some sort of charge.

And whether or
not gas a science that are normally

policed by some type of ion source.

And the coupling of ion sources to two
different must be common to techniques is

one of the main topics in my lab and
sometimes I and

sources are you know less than ideal.

So you can overload any given
ion source like the same only or

electricity very easily.

And by overloading I
mean that the response of

these Aryan sources is not
going to be represented.

Even if it's in the sample you're going to
get it in less than accurate picture it's

like having a digital camera with very low
resolution and that's a situation that you

really don't want to be if you're going
to be looking at a very complex system

you want to have the highest
resolution digital camera.

So the limiting factor it's not the camera
but your ability to look at the right

picture and so sometimes we do liquid
face separations and one of the most

popular liquid face abrasions is H B N C
that we'll probably seventy percent of you

use in the lab and then I'll show you
some experiments that use H B N C.

But even H B L C Sometimes it's not
sufficient to solve very complex samples.

Let's say let's take one of the most
complex challenges in systems and

then it takes it's looking
at human serum samples.

Hundreds of thousands of compounds.

It's not even you know even with these
very sophisticated instrumentation may not

be sufficient to to analyze
that type of sample so

we've been investigating some additional
They mentions of separation for

example doing separations in the gut
space after the I assert are formed.

OK And that's a that's a new topic in
my lab we've been working on that for

the last three years more or

less as many of you know we end
up with a massive data file.

And that's only data but contains very
little information at that point and

we really want to see what's relevant
information in that data and

that there are many ways of doing dad
that we normally apply some fairly simple

kinematic techniques and mostly for
selecting which features from these data

sets are relevant to answer
different questions.

So when my students do is
to put all this to work and

you know I just draw the nice lives.

So let's move on and let me let me tell
you what we're doing in terms of looking

at different types of foreign sources and
I know source are clearly

the bottleneck in the in these
type of problems they're not.

You know not to be underestimated
in We'd like to focus and

they be expensive instrumentation but
the bottleneck now is system and

it exists really look
different types of power.

And sources and that's been reflected for

many many years he must
become a true community and.

People have looked at the Ryan sources for
the last two decades and

that really tells you that this and
this is an important problem.

And you're probably most familiar
with Electress pray and Malia and

station those two techniques actually
were given the Nobel Prize and to create

an arc of formality and to your friend for
electors Brae in two thousand and

two because it was the first time that you
could look at proteins a large problem and

see the grass face without breaking them
apart and that really changed the field

in the last fifteen years we can do
things that we couldn't do before but

those are only do most famous topics
there are many other topics and

my lab is working on different things.

We're working on imaging techniques
we're working on doing Molly and

I will say pressure.

We're working on a whole new set of
techniques called and in the spectrometry.

So that we're working on U.I. and
sources it's clearly an important topic.

You may have heard of things like
Selby which is the role of the related

technique.

So this is a feel that
most very very fast.

If you look at the spacing
between these arrows.

It's actually not a linear line so
it's are created faster and faster and

faster and I keep updating the slide
every month on health basically So

the more the basic technique that most
labs use is called the Electress brain and

illustrates not a new phenomenon.

Although it was recognized with a Nobel
Prize you know you know two electors

waste known maybe four hundred years and
it's basically a process where you take

a liquid sample and you make it flow
through in the name of steel needle and

you apply high voltage and it's real and
when you apply high voltage.

These the end of the liquid dismissed
because basically the forms to produce

charge droplets that are generated
into the space between the needle and

they must become Mr and as
the distributor ejected this is call it.

Taylor crown here distributor
ejected they are forced

to flow by this electrical field
against the current of gas and

then they dissolve eight force by this
country a kind of gas they loose solvent.

And then the surface charge increases
because you're you keep the same charge on

a smaller surface area and that some point
that produces the droplets to explode and

produced smaller droplets and so on.

So once around finally produce
a diesel they did Iris.

And would you get into the musty
commentaries Aryans that

can be analyzed by spectrometry
it's a very elegant process but

by no means it's a very simple process and
believe me the mechanisms

are not completely fully understand yet
there's a lot of work but

there are some things that are not
agreed on there are many theories.

And you know this is really the
fundamentals of how electricity works but.

The way we do we like to spray in the lab
between very very simple instrumentation

and this is what I like to call
it before your morning coffee.

Electra's play experiment.

It's.

Not very difficult.

OK so you can do it like this pain fifteen
minutes you basically maybe like your

sample of this and maybe lights or and
on top of that maybe essential process you

apply high voltage you're charging to
those droplets and you're shooting

those charged troublous into the elite of
a mass spectrometer and this you know it.

Normally it's a heated capillary.

If he did so
you can increase the salvation and

what you get at the exit of
the scapula are normally aisles.

However some people seem to drink
coffee much earlier than our group so

in about three years ago M.

A group a producer with a cook's
group which has been leading the fuel

of my spectrometry for
maybe thirty years and

decided that how about doing
Electra sprain a different way.

How about Valentino doing it
the way we were doing it before and

shooting the droplets into the cutlery we
could maybe just point the Electress prey

at a solid sample and then those droplets
would basically bounce off your sample and

then dissolve part of that sample and

go into Electress brain in that way
you could do solids analysis without

having to you persist your sample in
a way you don't have to solve it.

You're doing.

It's like a miniature extraction basically
and that one point and this technique it's

called the suction Electra sparingly
sation So this is Electra spray and

it's a shame but it's a disruption process
what you're doing is you're basically.

Serving molecules from a solid
sample after the droplets flow

into in a spectrometer and we use this
technique quite everything in my lab and

I'll show you a different and different
applications but there are some papers

already out there trying to understand
in the any section make any sense.

How does this even method works
good Most agreed with a law or

process that people believe it's basically
that this droplets create a very thin

film of liquid on the sample and there is
one extraction so components have been

extracted into liquid solid extraction and
components are being extracted into this

film and then the droplets coming from
the spray are this large in this film and

part of those molecules on the liquid are
being traced into a mass spectrometer so

this is really an interesting paper in
chemical communications last year that

shows some simulations on how Dessie works
and has some some interesting results.

I'll show you some of our work also
using these type of simulations for

the Fed to find sources in a minute.

So our one of our first experiments
we would best see was to see what

we could do for themselves for proteomics
and we collaborated with a good money

manager North Carolina State and
we took our this line source.

And we started shooting at
proteins to pass it on plates.

So this was a very simple experiment this
was remember this is a new technology.

So you want to start
with simple questions.

So say you know can we see
proteins can we are nice.

Proteins if they are solid and deposit
it on a plate and the answer is yes.

And depending on the double
mastectomy that you.

You have on the side you can get
a very highly solution must look at

the best resolving power in
this case twenty thousand.

If it's so high that it allows allows
you to separate the different items in

the different isotopes.

So it's this is a very high
performance by spectrometer But

this same technique can be applied to
low power lower performance instruments.

So the nice of the nice idea is that all
these systems and extols are molecule and

you can combine them depending on
what you're planning on doing.

But the answer was yes we
can get very very good will.

In a very short period of time we
can restart in all three seconds and

you know we started dreaming what can you
do with this type of thing can we use it

to read micrograms can
we you know you name it.

So the possibilities are a really
open I mean there's a lot to do and

in this is this is very active field
you will see lots of papers and

conferences coming out on this there
was a watch up last year on this and

these topics and other things that we do
with this in my lab so M. and this is

one of those projects that basically you
get involved with even without thinking.

So this was a project that we said Sure
we'll just help you are good examples

you know in the beginning that
this was five years ago and

now it's become one of the the most
active projects in the lab and

what we do we use def see
this is there a photo of it.

Of Ed our first desk the source of this
with this was probably three years old now

and you are you see you
are shooting at something here and

this is something it's
a pharmaceutical tablet.

OK so you can take you know a solid
sample and sure some pull the surface or

work from that sample very quickly and

then the droplets bounce off the sample
and are collected by the vacuum and

the potential different going
into the mass spectrometer and

the particular service that we've been
looking at are anti-malarials And

the reason why and I didn't know of
this they have to admit is that there's

a very large market of fake drugs
out there and very very large in and

the scale is actually not very
well known because it's so

large and nobody has the the ability
of looking at this in detail but

this is a picture of a fake tablet and
you know in tablet and

you can see that the copies are not
perfect but they're pretty good.

So if you're buying these as a cheap
product and you really want to save money

you make in the brain the wrong wrong
medicine and what is important.

Well because drug quality really relates
to our ability to prevent disease.

So it's not only about diagnosing the
season understanding disease but also and

preventing disease and sometimes
preventing disease is just the cheapest

way to go about this rather than having
to diagnose a disease so M. it's not only

are the tablets are are counterfeit
socially packaging so they fake drugs.

Really come with fake holograms to and
one of us are very very very good.

I would dare you to come to our lab and
tell me which ones are fakes and

we transfer real they're extremely good so
this is not your cheap.

Fake anti-malarial he said
these are extremely good.

So you have to do some sort of
chemical analysis of you can for

example take your your suspicious tablet
put it in case they set up on a fake.

Will show sometimes nothing
sometimes some peaks and

a gentleman will show
the real active ingredient.

OK so this experiment takes two or
three seconds and the advantage of

these it's almost like a systems forensics
advantage where you have you can look at.

As many things in very large scale so

we get three hundred samples we
can run them in one morning.

OK we don't have to tell you.

OK.
Call me in two months and I'm going to be

done with the samples that sort of
response time is not acceptable anymore.

So we have lots of results on these but
I don't want to spend too much

time on this and we have focused on this
type of molecule very very actively and

this actually dose in a group at
Berkeley that is being synthesize

in these molecule using some you know
actually microorganisms for them.

This is the Artemisinin believe it
it's one of the latest generation

anti-malarials And

these are the only anti-malarials that are
effective in Southeast Asia at this point.

So this drugs it's you know we
don't have anything else except for

these drugs to treat malaria.

In those those regions of the planet so
presetting these many thing is

critical and the best way to generate
resistance against many things nice

to administer low quality drugs to
people so that's why the relationship

between preventing disease and looking at
low quality drugs is so critical and and

we can detect these of course
you seem decedent the problem.

I don't know if you can see it because
it's pretty low but the signal to noise

ratio in the spectrum is not so
good and there are lots of other picks.

Right.

What are those well those
other bits are fragments.

I mean what I told you there are a source
of should we serve the identity of

your sample not create artifacts and
in this case this he struck me.

One thing this molecule because
it's not stocked enough

that the survey from
process is pretty hard.

So we say saying well maybe we can
finally do some chemistry not so

much the signing of machines and
if you look at this

molecule it's very interesting because it
has this in the peroxide group here and

what we found is that if you take
a leading role kill I mean and

you bind you make a chemical reaction and
you've made bear with this and

with this group you get
a known prevailing complex and

that novel in companies actually much
more stable than the molecule itself.

And I apologize.

This is super This is light but

that's what happens when you have
strains that do a lot of work.

But what you get now is you get you
know these very intense signals for

only the complex of
a definitely different means.

So this is a way of the doing
a reaction in that this is prey so

it's a dynamic chemical reaction you put
in the Unlike you I mean in the spray it

reacts with the surface of
the sample its own sick complex and

it's injected directly into the into
the mass spectrometer in very it so

it's really one step process and we've
done extensive experiments with these but

I don't want to overwhelm
you with these so

other other things that we're working on
right now is looking at other drugs that

could be counterfeited in a global
scale and I wonder we're looking at

this this disc molecule here this is a
year which is an anti-retrovirals Tamiflu.

It was used for bird flu and as you
know many governments state governments

individuals have been stockpiling
Tamiflu because they were not

because scared of a potential
influenza pandemic.

So we said well can we do these
type of chemical reactions you

know that's the spray and
the answer is yes we can couple these

are certainly the molecule with
different crown eaters and

those Crown eaters react very
nicely to give you just one peek.

No fragmentation.

So the Senate.

No it's an additional very few cation off.

Is this the right molecule and
it's the right drug and

we've been recently when I say
recently I say office last night.

Ok this I got this from at midnight.

Working with the sherry group in K..

History tamales complexes.

OK And we're very excited because we've
come knowledge the stability of these

complexes and the theory seems to match
what we're seeing with the experiments

you guys face an insanity
experiments that So

this is a really project
ongoing right now.

So let me slightly go back now to
our systems biology questions.

What else can we do we said Well
high throughput measurements.

One one clear high throughput measurement
could make sure lots of samples and

those samples could be anything you want
the other high throughput measurement will

be the one sample by measuring
many positions of trust.

If you can so spatially we solve
measurements not not simply many samples

but one sample with many measurements and

we've been using this to
do imaging experiments.

So if you keep yourself break.

Fixed and

then you move your sample you can obtain
a mass spectrum from each position.

And then you can process each one of
us must pick try and you can plot.

You can make an image which is through
the color plot of the intensity

of the finance.

OK In this case where we did
clearly were looking for

seed funding when we did this
experiment we did the G.T.

logo and we do it with permanent
marker on a piece of paper.

It has to be digital right important if
you're going to be talking to me straight

to administration we do it with
permanent marker and then we imaged.

And it's a low resolution image we're very
quickly at one millimeter spacing because

we don't have any software to process it
data but it gives you a flavor of what can

be done and you can do a very nice imaging
experiment some people have shown for

example imaging of brain
slices with this technique.

Is So there's a lot to be done with
Jesse in terms of spatial measurements.

If you're interesting in looking at all
the groups involved in this technology go

to the October issue of the overall seven
issue of chemical engineering units and

you will see a lot of the best C.
and what they do with their C. and

some other techniques this it's only one
of these group of techniques called and

been much spectrometry It's called them
in because it happens in the open air.

No enclosures it's really disappointing
nothing that prevents you from looking at

very different samples.

In this actually picture in our lab.

We're looking at three hundred miles with
a different technique this is a thing

we call the DART It's a now it's a
commercial name and we're working on right

now it's really understanding how this
type of blast my base techniques work and

you'll see down here at the same movie of
the table simulations that we're doing.

Just to show you know how this
step of the structure things work

how they heat out and so on.

So I don't have time to show you all the
details about this just want to give you

a flavor of how fast this
field is moving we can

also in some simulations to understand if
you put some for some different places.

How sensitive these methods
are OK You see for example it.

If you put it right in
front of this aisle source.

Most of your material is blown away.

This is the must become a very light.

Doesn't work.

OK So more complicated and
it seems OK getting the best sensitivity

that's why I never say Sure I mean the
samples will have been ready after lunch.

Some things are challenging.

OK so.

I think I'd bombard you with things or
die and sources.

So what else do we do well.

I tell you that we do a lot of
work with separation Thank you.

Metrics and one of the sensor we are
heavily involved with the McDonnell lab

and really working very hard.

Is to find a way of using these type
of high throughput approaches for

medical diagnoses and
that we've been working very heavily and

trying to find in metabolic diagnostic
pattern for ovarian cancer.

So what we're really looking at is not
a protein biomarkers that's something

that probably you know a hundred groups
working on but looking at metabolites and

the reason why we think that Metabolife
are a better way of the notion of

an cancer is because first of all
they're easier to detect that's a key

point go after what's easy to
detect in lower concentrations and

we work in about these for
a year they need to see if a person

is released early stages of ovarian cancer
we have a different metabolic pattern.

Can you measure all
the metabolites at the same time.

And then tell me if these parents that
you see are different and you would see.

For a person it's healthy.

So it's a very very difficult
question as you know.

And the process.

It's fairly complicated from
the experimental point of view but

it basically boils down to taking
assume sample and some if you some

preparation steps in this case we can get
rid of any sample preparation steps and

doing a little chromatography separation
and then followed up by much spectrometry.

And this is a special type of mustard
trauma too which is highly solution

must be trauma tree so it's possible
you can identify metabolites.

And then we do a lot of camel metrics on
these datasets to see which differences

are just biological and which differences
are really caused by the seas and.

The amount of data used by the sighted
with metal since it's extremely high.

So this is just an eight this is a single
H B N C run in a single H B N C When

you can take up to.

You know three thousand metabolites OK and

my student is the pasted the H B L C
chromatograph and it was this long.

You know it's a three hour long
chromatograph and this is for

each peak you're looking at
the mass spec to for each peak and

this is I don't think I'm going to go into
this but this is a type of the you get.

It.

It's basically delineation all day
that we can see hyphenate in this.

And you get retention time on this our
system as to church and the success but

look at how rich.

This is just one serious sample how rich
this is there's so much information.

And the question is Are any
of these metabolites under or

over explain what's going on.

And as you can see we don't
know the identity and

we're just talking about
peaks at this point.

OK so the second challenge is OK does
Pixie's differentially you know.

Express the detective.

Is this relevant to the seasonality
I want it isn't our lives so we

heavily rely on many can all metrics tools
to do this so we take this broad data.

And we train we train the genetic
algorithms to select which

peaks are actually important and

which bits are not so important and
when I say important these can be said.

Subset of peaks from the crowd around

differentiate between a controlled
person and the animal and

the patient with ovarian cancer or
not and we let the computer try these and

you know thousands and thousands
of times until the classification

are basically slowly goes down
until you end up with a subset of.

Of peaks in these chromatically on that
it's much better for diagnosing cancer and

let me show some results and
these results are already old M..

But this is the first sample said that we
tried this about twenty four samples and

that we were trying to see if we could
by this approach you could distinguish

between ovarian cancer and control samples
and this is a decision stressful here and

all the samples are ovarian cancer serum
samples and these are all controls us.

And as you see there no
misclassification no false positives or

false negatives happening.

And so this is very promising.

So after we did this maybe three months
ago now we just finished a study

with seventy samples and
we get the same performance.

So it's a very promising tool really
to look at the look at your diagnosis

and disease.

What's not good about this
is that it's not very fast.

OK it's already a two or
three hour long H.P.L.C. run.

So it's less than ideal in that sense.

So there's still challenges but
you know we're making progress and I don't

know if you notice when I was
showing you this blog here and

basically this part of it came out of I
was completely saturated you have a lot

of the from molecules that are basically
looking at the same retention time.

That's telling you that
even though we wait for

three hours we use the best column.

There is separation it's not perfect so
you only thing that you can do is you

could do for example a Bill
mentioned chromatography right.

Add a second axis of Commodore appeal it's
a strong gardyn exchange people do that in

person or makes or but the problem with
that is that it makes things much slower

admission and chromatography
experiment takes about ten hours or

you could use what we use which is
a gasp a separation technique which is.

Happens in English take
on talent scale and

this is a technique that is
called mobility spectrometry and

it's place basically after the iron
stores so molecules separating

the chromatographic crown him are iron
eyes they further separate into this.

Drift chamber I'm going
to show you in a second.

And then they're mass analyzed and
then with that you really need very

very complex algorithms to distinguish
was important in this data.

So I wanted to you know to face the dark.

I wanted to show you what we're doing
in terms of glass face separations.

And I am ability to actually believe it or

not it's the most popular
analytical technique.

There is there are fifty thousand I'm
not released I'm an event deployed in

the world every time you go through
to the airport there are two or

three hour immobility instruments
sitting right there and

I am already dismissed they started
you know basically been developed for

chemical warfare lead to
action in there in the field.

So that's the way people expect me to
workfare biological warfare and you know

you can help portable units handheld
units and you may have gone through

one of these if you have a strong accent
like I do I always go through this portal.

Right.

Just in case they put the price that they
work in a lab that has a lot of chemicals

so sometimes they do
get some positives but

what's basically happened
if you go through this.

Portal they blow away any part because
that you may have on your clothes and

those particles go into an I am really
spectrometer OK and they are going to

spectrometer if it's small enough that you
can have it in there get bandages that you

don't need that there is no need for
a vacuum pump that you have to maintain.

OK there are some I mean what we did
mention is that you struck him but this

issue is on your show you hear all and
atmospheric pressure separations and there

are some new technology and I am not going
to be that allows you to separate viruses.

For example if you want to separate
viruses you can't you can separate make

adult particles you can using a movie
spectrometer There's a company

in Minnesota called the side that makes
these beautiful instrumentals for

separating micro Mark Ryans
that's what they call it but.

In how this our immobility work.

Remember this is a separation stage
that we're going to put in between and

now our own source and
our master dramaturg and

this is basically you know you have to
have some sort of I A source in this case

it's an electro spray and
such but it could be anything.

Could be a mole the source.

And then I and so
injected into this drift you.

And there's a potential
gradient going down here so

I use our force to flow this way.

But also you inject in
counterflow inject the drift gas.

So then there's a discourse of the effect
where Alice are you know slow down.

Depending on their shape.

Not just their mass.

It's like a glass vase
electrophoresis basically So

what you're doing is you're separating
things based on Mars mass but

also based on size and
that's a unique advantage because remember

that it must become a there won't be
able to distinguish the nice summers.

So if you want distinguishing
isomers let's say Guy counts.

This is the type of tool that
you really want to use and

then you have to have some sort
of the text or the simplest for

it sort of It Takes A It's just the plate
make sure the current of the eye

is killing the plate but
you could put in a spectrum of the here.

OK And then you will get separation and
then must become any detection.

This is what we call a stand alone and
mobility separation and

I am not really has a huge
history at Georgia Tech.

This is the first one of the first
group cells ever reported and

this was this is very humbling because
I was negative three years old.

This is the judges of knowledge in
March of one nine hundred sixty seven

physically parliament and.

And they reported this first design for
how to make a name mobility instrument and

basically starting actually what's
being used for the last thirty years.

OK So there's a little bit of a garbage
like a forty year gap since this and

we started working but very interesting
that this happened on the street.

OK so sometimes you know.

It's very high and
it depends on the pressure it's about in.

Four hundred volts percent meter.

You know that depends on the pressure you
explained about from a straight question.

It's very high because you're really
pushing your ions to flow through

a very thick guys that said
that's an important point.

So this is the one that we didn't.

And we said well can we do something
different or are we going to do the one

thousand nine hundred seventy thousand and
one of the problem of this design.

You probably notice that
there are different rings and

those rings are metal rings connected by
resistors to generate a gradient downhill.

By the grain is then going to be perfectly
going to be more like a stair step right

with a sign.

So what we decided to do is we started to
say well can we use a new materials for

these and there's a company that at the
same point was was basically studying to

to manufacture this type of monolithic
drift to use this is the glass tube and

it's a glass tube that has
resisted coating in it so

you can apply voltage at the beginning and
at the end and you have a perfectly

homogeneous electrical field downstream
ready to tell you it's a glass tube.

But it's conductive more semiconductor.

So on and but the rest is basically
the same it's a nano spray ion source here

in a little dissertation chamber
to get rid of the solvent and

then annihilate ions
when you get to this day.

You open and close is great and
then you let them separate.

And when asked they which they
take to you measure the current or

you measure of the mass
with a mass spectrometer so

this is very similar to time of flight
experiment that you do must pick

trauma to me with the difference that this
is done and the atmospheric pressure.

So there's an additional separation force
which is that is constantly of these guys

this is really not much spectrometry but
it's used in combination with mass spec.

So this is going to be a long time
to be a lot of collaboration with

the streets company and you know we're
just now showing our first results but

to give you a flavor of what
the separation looks like it's

if you like I don't show
you this axis here.

This could be an H.P. and see trace
right looks very similar big big big.

Nothing new.

What's very different is that.

The axis here it's in
the millisecond time scale.

So you can do your hour long H.P.L.C.

separation and
then you have another dimension.

It's admitted they mentioned where you can
separate things in the millisecond time

scale and the mass spectrometer
works in the microsecond time scale.

So it needs milliseconds microseconds
all the things can go work in concert.

Then you can get
the multi-dimensional separation.

So M..

And this is just a very simple molecule I
mean we're using research been researching

with our hydrogen molecule it's
the simplest molecule we use from aspect.

We do everything we said it's just a drug
molecule nothing interesting about it but.

This is not a perfect technique and the
problem with the spectra here is that most

of the sample has been lost because
of the way this process work.

Works.

So I tell you a young audience comes
to come to the gate you open and

close the gate and you open and

close it was a very short period of
time about four hundred microseconds.

And then you wait.

And you let the audience fly
towards the detector they fly

four hundred milliseconds.

And then we're not they all reach
they take their you open and

you close the gate again.

But while you're waiting in
here all the Iranians are lost.

So you spend four or four years purifying
the protein and then we lose ninety nine

point nine percent of the ion so
the ion source is not acceptable.

It works except all cells were
trying to they were trying to

improve the duty cycle of the system and
that means the percentage of

usage of the ions How do you do that well
open and close the gate much faster.

Right and the type of gating
sequences that where it is

starting to use now are basically decent.

Fully random sequences.

But they're not random.

They look random Your opening and
closing these gave me a random fashion but

you know exactly what the sequence
looks like and then you detect

a signal at the end and you do
a deconvolution process of the signal and

you can read your spectrum so this is an
approach that for example is used in women

microscopy Romanise in very low
intensities in phenomenon strong you

have to use some sort of
multiplexing technique to increase.

This signals.

Same thing in a mobility and people for
example that Pacific Northwest.

National Labs are doing this from
a spectrometry very very interesting and

hard topic as well trying to work
as fast as we can on this matter

just to show you what you discuss
to offer in the blue line is

basically a low they loot sample running
the conventional mode where you open and

you close your gate and
you wait the red line behind it.

It's the multiplex experiment and

you can clearly see that there's
again a signal to noise ratio.

So we're working on these
type of things right now and

i already has more to offer
than just a separation.

I told it can actually separate isomers
And I think that a challenge from

a spectrum of trees to move from you know
it wasn't it moves they say from atomic

mass spectrometry to molecular and then
to structural aspects monitoring how we

get structural information and
you can I am mobility the mobility of

Finighan actually depends also on
the cross-section of the ion so

you can separate things based on shape and
you can calculate I am or will it

is one of the only ways of calculating
it an experimental crocks cross-section.

So just to show an experiment
that where we did.

If you take a complex What's
this in Iran complex and

we're looking at this see there
are four and complex this and that.

And by separating them by immobility and

if you look at their league and
complex with the middle.

You get to pick here.

But if you look at the free legal.

And the full and

it's actually lighter than the complex
legal you see it big coming much later and

it's actually wider and that's because the
feeling on cross a larger cross-section.

And so it takes longer to reach the end
although demand is actually lowered.

So these numbers actually explain mental
numbers you can match those experimental

cross sections to of molecular
dynamic monitoring to

try to understand the confirmation of
this or you can in the US face him.

So there's much more to him but we don't
know just a separation they mentioned so

I wanted to give you a flavor of that and
with that I'm pretty much done.

We have a new.

It's not the Web site that
Jeff was referring to but

you know we put up a new website
a few weeks a rower we can go on.

Of course I'd like to thank my
research group they do all this work.

I just make the nice Powerpoint slides.

Thank you.

Of us.

You know.

They're going.

Yeah.

You are.

So worse here but.

Here's the deal you know this

up to the optimization stage of the system
is essentially they reach information and

that's exactly the type
of experience that we do.

There so we said balance in sensitivity
versus resolution in something someplace

being versus you know
spatial resolution and so on.

So you can make the survey slow response
separation engagement very high.

We saw in power or you can make a very
fast and just get partial resolution if

you're going to have a high reception
power must pick at the end.

You know you need a partial separation so
many things to gain there.

But yeah very good point.

And we've done this experience
a different flow rates and

you see the eyes moving back and
forth as you would expect.

Work every

single year.

We don't know the peaks.

At this point we I mean we know
some of them they're mostly bits.

I mean the mass range.

It's up to two thousand
more like in a more C.

But most of the relevant ones actually do.

No five hundred.

So these are really small molecules and
we have identified maybe seventy

percent of them using our good mass
measurements and that our research is and.

I would say that most of
them are some sort of leap.

You know and those are very abundant
too that's why we're only taking them.

You can always big deeper
into this type of signal so

you have your work sitting thresholds to
how many books you want to detect and

things it's very complicated when you're
looking at the big race and on signals but

even with these major components and
this is the pattern of our twenty peaks.

We seem to get very good
miscommunication between the groups.

Then we see finders on your question.

You see the different
stages of a brain cancer.

The X. excellent point.

The eight emails me constantly
are these type of things and

I I appreciate that there are questions
we're really scratching the surface.

OK So the first crude experiment
did say ovarian cancer control.

But as you know that's not sufficient.

You have two different stages of cancer.

So we would like to see if there is also a
possibility of doing this type of morning

where you have three or four classes and
distinguish between that and

the data will be the same but
the challenge.

You know it's much more challenging
to distinguish smaller differences.

I can I CAN WE ARE VERY SOON
some effects that in

those cluster plots some of the earlier
in cancer samples actually cluster

slightly different than the other ones.

The problem is that you don't
want to make it generalize.

You know assumption that these were
trying to happen it's very hard to

get really ovarian cancer samples.

And then you can make a generalized you
know conclusion because you don't have

any samples and then you don't have any
samples because you don't have a method to

detect them but you can detect the method
because you know how the early samples.

If that makes any sense but
you see what I'm saying it's.

A catch twenty two for the early
summer is that we did the detective

we can distinguish them from a normal
IP because we're pulling them apart.

We gather a number of late stage
samples we don't know yet but

hopefully you know maybe in
a year I can answer that.

Users will do so

because they're all of you or
will say no it's not the same no.

So if you just figure something you do
have something simple like a P.C.A.

you see sub clusters and the point is
that each patient has said they reach

clinical history and
to which extent that legal history

relates to the class about your scene it's
very hard to tell even we know we have.

Almost eighty samples now and

it's hard to really know if
those clusters are real or not.

And but we do see some clustering and

I'm very careful about drawing
any biological conclusions

from them at this point but I think that
if you kept increasing the sampling.

You would you would see more
important effects results

would say you know we
would love to do that

we would love to do that and
some of the one of

the problems with that is that you're
always sample limited in the amount of.

A tissue that you get and that reflects
on how sensitive your measurements are so

we use normally something like four
hundred micro leaders of serum for

doing one of these measurements and.

I'm not so sure we can do is make the same
type of measurements just from tissue in.

With the current instrumentation
that we have in the lab and

you may need something more sensitive
to rules but it's very important.

You know to that type of correlation.

Frank I know you have a question.

I think yes.