Tip:
Highlight text to annotate it
X
If everyone could take their seats, we will start up again. Are the three of you going
to scan the whole time? You look like you are ready to be shot over there. We are doing
these one at a time. The two people who are speaking don't have to stand, but we are now
going to shift gears a bit. Panel three is on disruptive technology in the day-to-day
loose. The first speaker is Jeff talking about challenges of the changing genomic landscape,
advances in sequencing technology.
Thank you, Eric. Many of you have walked around in laboratories and you know that laboratories
have benches and water bottles and if you interest -- instruments scattered around.
In contrast, the laboratories that sequence the human genome for the project tend to look
more like factories like this where you will see rows and rows of seeing good thing machines.
This is what it looks like when the human genome project was being completed. During
the time the genome project was done, the cost of sequencing came down quite considerably.
This is the result of several things, technology improvements, automation, and economies of
scale. We noticed that it was taking about 10 years to reduce costs. At the time that
the human genome project -- was a earache or Francis said, it costs about 400 million
or so dollars. If you had done it again at that last moment, it would have cost about
$50 million in about 2003. As you know, what we are here -- part of what we are talking
about today is the subsequent technology developments that have brought those costs down considerably,
which now seems like second nature, but during the planning process -- actually, I guess
no one has actually mentioned the planning process. Eric will talk about the planets
process that we are embarked on right now, but the last time we did that was in 2002.
People asked whether there might be technologically difficulties. There it seems so far off. If
they could be achieved, it would revolutionize biomedical research and clinical practice.
One of the several ideas that was listed there was, do you think it might be possible to
develop technologies with which the sequencing [ indiscernible ] $4000 or less? Fast-forward
to today. We are not there yet, but we are actually getting reasonably close. This slide
just represents one of the reasons we got there. That is not the whole workflow to doing
sequencing has changed dramatically. So, the process by which we sequence human genomes
involves not only the sequencing machines but lots of robots, large machines, upstream
of the sequencing machines. Handling every single sample in a separate test tube. In
contrast, today we can make a library of a whole genome in one test tube and apply that
to the sequencing device. Second, what we call a run. That is running the machine once
on a bunch of samples. You could only analyze a sample at a time. Now the machines analyze
something on the order of 100 million samples per run. They take about a week or two where
as they used to take an hour or two, but even so, the multiplication -- the multiplier there
is very dramatic. The process is simplified. So, what does it look like? This is today's
sequencing machine look at. There is a microscope in there, essentially. They see the little
dots. Each dot is a feature that originated from a small fragment of the human genome.
But the machine is watching is as time progresses over minutes or hours, it is looking at the
change in color of each of those features. Just to give you an idea of the scale here,
100 microns in the lower left is about the size of a human hair. That is how many features
you are looking at. Of course, the slide extends vertically a considerable distance. That is
the number of different sequencing reactions that you are looking at in the size of a human
hair. So, I launched a program to advance this technology development. I won't really
talk about that program in this talk, but you have some reasons about in the materials
that Larry thank you links to. So, some of the outcome of that is the series of commercial
machines that have been released starting in 2005. Many of these machines have been
upgraded now and new machines keep coming out every year to produce these torrents of
sequence data that you have been seeing. I am not going to go through this slide in detail.
The point here is to say that there are quite a number of machines that are out there now
doing sequencing. They have considerably different features. That is good for any number of reasons.
That is the quality of data you get from different machines. Different machines are good for
different experiments. Perhaps most importantly, there is competition. That is really valuable
for accelerating the pace of technology improvements and accelerating the pace of cost decreased.
Another advantage of these machines over the machines that were used for the sequencing
of the human genome is they can do a number of different kinds of experiments. They do
the experiments in different ways. Again, I don't have time to go through that in detail
today, but a number of the kinds of experiments that were used for -- that were done by micro
arrays are done on sequencing machines today. They give us more quantitative data than we
could get on the micro overlays. These are very versatile machines. To give you an idea
of the kind of technology advances, I have prepared the slides. If in 2003 you calculated
what it took to sequence the human genome, you can run through the details here, but
the bottom line is with the technology, it would take about three months with 100 machines
to sequence the human genome. I do that calculation in 2007, so that is shortly after the first
few next-generation machines came out. It would take about the same time, three or four
months, but instead of 100 machines, one machine. Late last year, the same calculation now down
to less than one month with one machine. Projected by the end of this year, one week with one
machine. But you saw all those little features. You can now pack so many of those genomic
fragments on one machine that in one week on one machine, you will sequence to genomes.
So, the numbers are just stunning. I had to comply with the prediction that someone was
going to show the slide again, so there it is. This was updated for another year. The
costs have come down considerably. Where are we going? We are not done. The technologies
we have in place now are producing vast quantities of reasonably high quality data, but there
are a number of challenges to using them. Several of which you will hear about in the
next couple of talks, but we are not done with technology development. There are new
machines and new approaches on the horizon. Some of them eliminate the microscopy and
eliminate the custom reagents. This is a company that announced earlier this year the availability
of its machines. This is instead of using fluorescence, which most of the other machines
use, this simply detects the chemical change, the change in pH, the ion concentration. The
earlier ones have one and a half million pH on the little chip. It is this big. It's a
little computer chip. It would be a purely electronic read out. Other approaches, there
is an approach that uses what some people call free running [ indiscernible ] being
developed by two different companies. Both of them use nanotechnology as a key feature
in enabling the visualization of the sequencing process. Again, because of time, I cannot
go through how the technology works. The bottom line is that this technology cannot only read
off the sequence data because the trace on the bottom shows, but very importantly considering
what you just heard in its last session, it can directly read from genomic DNA models
without doing all the chemical conversions that and he talked about that they have to
do today. They are quite expensive. It can directly read from genomic sequence. Where
is the pointer? There it is. This is a big advantage if this technology end up working
extremely cost effectively and at very high accuracy. The other thing that these technologies
will probably offer is very long life. A technology the first time gave approximately 800 based
read links. Most technologies today get somewhere between 100 and 300 or 400 based links. This
is a big deal when you are looking for some of these other kinds of variations in genome,
such as [ indiscernible ]. It is very hard to get with sort -- short reach. If you get
long, it you can put them together much more easily. Finally, what a number of people are working
on sequencing technology. The idea here is that you will put a DNA molecule through a
small hole that is the dynamic of a molecule. If it works, if you flow ions through the
channel and the DNA is going through the channel, the different bases in the DNA will disrupt
the ion flow differentially so you would be able to read a versus G. versus E. with a
purely electronic signal. Again, the advantages here -- let me just show you this and I will
talk about advantages. Once again, people that have shown you cannot only read off individual
but also methylated bases using its technology. Once again, no conversions. You could do these
measures directly isolated from cells. Potentially on very long molecules. There is a long list
here of advantages if this technology works. Long reads, working directly from genomic
DNA. It is not destructive, so you could go back and read the sample over and over again.
This would simplify a number of experiments, for example microbiology sequencing. If this
works for a DNA, it should also work for our DNA -- are and eight. -- RNA. It would be
a fully electronic method, which means you would be able to use small handheld devices.
They should be deployable in other than large sequencing centers. This is a tremendous interest
not only to the genome sequencing community but to several other communities like homeland
security, for example. For monitoring emerging infectious diseases or potentially bioengineered
disease agents. I will wrap up with this last slide to say if this approach works, how long
would it take to sequence genomes? With a small array, an array of just a thousand nano
ports, if you could do this, you could make many more. It would take much less than a
day. Perhaps just a couple hours to sequence a genome. Okay. Our second speaker is --
this next one down? -- Vivien Bonazzi who will speak on data deluge and analysis. I
think just give you some numbers, but I want to give you some context here, especially
the old sequencing methods, which are really not that old, generate about 30 megabytes
per run. I am just using averages here. But when you look at the new data volume, Jeff
also had a slide that cover these numbers. Depending on how you do these experiments,
you can see just from the numbers, the volume is tenfold, 100 come even greater. As these
new technologies and third-generation technologies come out, I'm not entirely sure what the data
volumes are going to be, but we know they are in the [ indiscernible ] ranges. That
creates a huge problem. Or they give you an example here. Most of what we have talked
about, and I assume we talked about it this morning, describes race [ indiscernible ].
It is not a one-to-one ratio here. The ratio is one to 20. These numbers change depending
on what you actually store and based on the technology we have. I would like to. out that
a base is not a bite. You have to think about what it is you want to store. If it is one
to 20 or greater ratio, the problem is, when you store it, you have a a lot more to store.
What is the cost to store this? This has implications for data centers and how you store it and
what you store. For the computational challenges, remember, informatics is a combination of
biology and computing. You have to match those two fields. Clearly, you are going to have
an infrastructure that needs to be able to handle this type of data. In the past, as
I showed you before, when you have stuff coming out of this Cap, most biologists who had training
like myself can usually write programs. As you start generating data come you need to
bring in folks with IT backgrounds and software engineers. Just to summarize some of the computational
challenges to support this kind of work is the data storage. As I mentioned, with the
volume of data, it is going to impact the storage. Secondly, when you want to do something
with that data, it will impact the type of analysis you want. The CPU would have an impact
here as well. It would not be able to run this on Excel and your desktop. You can try,
but it will die. The next thing here is we need to develop new hardware for this kind
of work and also software. I will explain a little bit about those but we need different
kinds of architecture for this data analysis. When you talk about volumes of data, I know
you have spoken about this in a previous talk. A lot of these next-generation sequencing
machines are going to be covered and placed in labs that generally didn't do this in the
past, which means you have large sequencing centers. But you've got to think about the
smaller labs doing this as well. Now you have a large, Jane Norman is sequencing centers
and many smaller groups who are now going to have one or two machines in place. I think
one of the things that just said was about 70% of the machines [ indiscernible ] con
is that right? Think about that. That is now going to impact the amount of data. When you
want to do something with that data, you've got to move that data around. That means it
is going to impact the way you can transfer data. If you have a lot of data, most of you
who read use the Internet realize it can be slow oftentimes. At the same time, we have
[ indiscernible ] to put in file. Why? They are actually increasing the pipes, but it
takes time to do that. What is the impact of doing this on this kind of work? Last but
not least is data security. Obviously, as we move data around, particularly dealing
with human subjects data, how do we protect that? This is really strong about how we make
sure to generate this data, we are very careful about not releasing it without such policies.
Here is our challenge. We've got to list data. We are a assuming we have looked at some of
the infrastructure pieces. In the past as I mentioned on the top slide here, you can
drink from a water tap, but if you tried doing it from a fire hose, you are not going to
have a lot of success. If we look at the analysis challenges of assuming that we are working
on our infrastructure challenges, we have to look at developing new tools. We've got
to have not just more data but the ability to compare complete genomes against each other
is just something we've got to do. This new type of analysis we can do, we need to be
able to do that. We need to respect her old tools important for doing analysis. For example,
[ indiscernible ] energy in finding tool. Both kinds of things need to be [ indiscernible
]. They don't necessarily run on the system in terms of the infrastructure or they don't
work because the sequences are very short. They are very short reads. We also need to
optimize a lot of these tools to work on the new platforms. The other thing is I think
we need to have improved visualization methods. As we generate all of this data, computers
are going to die and we visually want to see things in graphs. We also have to think, and
I will bring this point home again. As we generate data, think about [ indiscernible
] that don't have high [ indiscernible ]. They need to be able to crunch this data,
see it efficiently. We need to realize we are no longer catering to those geeks, which
I include myself, to people that can actually use this. But to say the same point again.
Robust tools. Making those tools robust often when we write these tools, it doesn't transport
while two different systems. We need to make them robust particularly for the non- [ indiscernible
] specialists. Data integration is also really important. We have all this volume and data
and we want to analyze it. We need to think about integrating it with other data types.
For example, we've talked a lot about genomic data. Clearly, functional information is extremely
important. Integrating that will be key to looking at function alongside. Included in
that, obviously, is a lot of image data, out. As we add this image data is quite large.
You have to think about the compression. You also need to think about that kind of data
as very valuable in terms of visualizing what you can see from genomes. How do you integrate
those two pieces? I won't go into a lot of detail, but another piece is the [ indiscernible
]. Often it is related to some sort of experiment you are actually doing. You want to be able
to capture the information about that particular experiment. For example, the run, run length.
Any of the experimental information. You also want to link it to potentially clinical information.
For example, we have had the micro Piatt -- micro [ indiscernible ]. It is all related
to human and you want to link that information back. You've got your sequence information
and you want to take it back to the actual patient itself. Think about what that data
is. It thinks of data volume, how you store it, clinical information that you need to
protect. You need to think about these pretty carefully. Last but not least, as we generate
this data, when you have a lot of data, you need to think about standards. If you don't,
we will have to disable. We have certainly had examples of that. It would be [ indiscernible
] in honestly. What are we looking at in terms of solutions? I am not claiming that
we've solved it, but we are thinking about it. This is data reduction. There are bits
of [ indiscernible ]. There is just bits of data. There is discussions that relate to,
do we keep the derived data? For example, do we keep the assemblies? We don't keep the
world sequences. We keep enough [ indiscernible ]? We don't have to keep all of those terabytes
of data. Do we keep the genes listed? Do we believe the assembly is correct? As these
new technologies come about and they become cheaper, you have to remember you can resequenced
something rather than storing it. It might be good to generate the base pairs and throw
away the rest. If you make a mistake, you can sequence [ indiscernible ]. What I forgot
to mention before, everyone had these lovely slides that talk about reduction costs. I'd
like to remind you that the cost of storing something is actually no more expensive than
it is to actually sequence it. That is part of the problem. The twofold issue here is
it costs more to store it and you have a lot more of it. It is becoming much more of bottleneck
than it did in the past. For reasons I described before, we need to address those. A couple
additional things is we need to actively involve the community in our discussions. We have
done that. In this case, both the biological community as well as the computing community.
We have had two workshops recently. There are some folks I've talked to recently. One
was the informatics [ indiscernible ] which talks to the community about the analysis
needs, the kinds of things we need to solve. It sits that particular workshop into the
overall NHGRI planning purpose, which I believe Eric has been talking about. Also, the cloud
computing workshop, which I will explain a little bit. And the last piece here is education.
That is, we need to be able to train people in the use of these tools can't train more
people to actually develop these tools and be able to integrate those pieces. And the
last slide. In terms of the solution, informatics is both biology and computing. The top of
this slide is really that integration piece. I didn't speak much about it today, but one
of the possible solutions we could look at is a way of storing and managing and doing
the analysis of this data. I haven't gone into it today because we just don't have that
kind of time. Those of you that use the cloud example, a lot of people are using it because
of the ability to store large amounts of data. You can actually do computers within the cloud
which speeds a lot of things up. The key here is we really need to integrate and think about
how we want to do it. I am going to finish with this last slide. This is how I feel a
lot of the time. It really highlights the problems I see in terms of biology and computing.
The only data volume issues that we have issues that we can't talk to each other very easily.
Quite frequent the I am asked about the database. Yes, I tell them, what color do you want?
We need to speak about how we communicate with each other. Even if we solve these problems
in terms of analysis and infrastructure, we need to solve the problem of communication.
I guess that's part of the reason for being here today. Thank you. [ applause ] Okay.
And our last speaker in his panel is Jim Mullikin who will give us a view from the sequencing
Center director. I have no slides. I was asked to give this talk at an accelerated pace just
like sequencing is running right now, since we are just before lunch. I will try to get
through what I was going to talk about today in a little west time than I'd planned. Previous
speakers introduced already both what is coming in the technology world and maybe what the
solutions are, but I am running a sequencing Center now that Eric ran up until just December
of last year, which he initiated back in 1997. When he started up a sequencing center back
then through cooperative agreements through all of the institutes because he really felt
that having a sequencing Center was important, he established the NI age intramural sequencing
Center. He was able to hire six people and purchase at the time six of the latest sequencing
machines. Things have come a long ways since then. We now have a budget of $7 million a
year. We have 42 people on our staff. It is a wide range of expertise. A lot of robotic
information. The laboratory we have is about 5000 square feet in size. With the rest of
the space, another 5000 square feet or so set up for doing all the analysis and storage
of data. Prior to the introduction of the sequencing technologies, we were operating
at a level of about 7 million sequencing reads per year, which is quite a nice pace for medium
scale sequencing Center at which we have. We were working on projects like you for today.
[ indiscernible ]. How has this dealt with the latest round of disruptive sequencing
technologies that have been introduced? -- and the associated daily data used. First
of all, we do keep our eyes open. We know what is coming in the technology field and
try to be prepared for it. We also work closely with other sequencing centers, letter centers,
like the Brody Institute and the Washington University genome sequencing Center. They
have enough capacity to really try out the latest technologies at the very early stages.
So, we do take a somewhat cautious approach. We listened from them and then once we have
an idea of what will be the best type of technology to bring in, we will bring it in any R. and
D. type environment. [ indiscernible ] do that with one of the machines, what was then
called the sequencing machine is now owned by Lumina. About a year ago after it had been
there for a about a year, we realized it was time to move it to production. It was transitioned
into production very quickly. We now have two kinds of sequencing machines. One is the
fourth by a four and the one other is the to ask. -- 454. That is five orders of magnitude
more than in 1997. We have come a long way. It has caused some stress. It has caused a
lot of strains but we are also quite excited about having this new technology. Just to
give you an idea, the lab had to learn new protocols. Technicians learn how to operate
the machines. The sample tracking system had to be reprogrammed. The number and diversity
of kinds of projects expanded, data transfer rates changed dramatically. The software development
team was stretched to its limits at times trying to keep up with not only processing
the data that was relentlessly pouring through because these machines were online but we
also had to adapt every month or every two months to a new software update from the sequencing
technologies to release new data with those machines. It causes great stress, but I must
say it is great potential as well. Be intramural program is invested in a lot of the expansion
of the sequencing machines and related infrastructure. Now we can support various types of data in
our processing so we are working with [ indiscernible ] type experiments from investigator driven
projects like chips seek experiments. It investigates interaction between the proteins and DNA in
a cell. [ indiscernible ] sequencing looks at Dean's in the nucleus of the cell and medical
sequencing that looks at changes in the DNA that are associated with an individual's cause
of disease. We have a mix of projects at the sequencing Center. We have a large scale projects
like the human microbiota project and diagnose disease program and [ indiscernible ] which
you will hear more about later today. It is great to have in a production environment,
because that gives us kind of a buffer of a whole bunch of samples to keep processing
as we move forward. Also, it accepts into the sequencing center dozens of smaller projects
that will range in size from one to two samples up to tens of samples. Over the last two years,
we have been bringing in these kinds of projects, the small-scale projects, at about one or
two per week. It is an incredible rate of induction of these projects. We talked to
each of the investigators to figure out what they want to do and implement an optimal sequencing
type approach for their research. To give you an idea, we have already had 20 of the
smaller scale projects completed. We have another 33 of those active in our pipeline.
Another 14 are waiting on delivery of samples. One of the issues with receiving samples,
especially human samples, is that they have to be consented properly before we can apply
some of the sequencing technologies we're using. For example, [ indiscernible ]. They
need to go back and review their protocols to see if it is okay to have a simple sequenced.
If not, they need to go back and get it re- consented. Once we do have them and we run
them through the sequencing pipelines, we can give them variation information very quickly
within just a few months of delivery of the samples. Having had the next 10 sequencing
in production and applying it to various projects, some of the first publications are starting
[ inaudible ] through. One is titled sequencing of axons. It causes a syndromic form of [
indiscernible ]. Senior author on that paper. That is just the tip of the iceberg. There
are many more projects coming through that are transitioning into the clinical -- well,
we have been in the clinical focus. Now with the sequencing, we can really push that forward
even faster. And at a scale that the individual investigator can come to us with a few samples
and with their own budget have been sequenced and [ indiscernible ]. It has opened a whole
new world. Even though it has been a preacher Moss was transitioned, it has been well worth
the effort. I think a lot of good science is underway. We will continue to have that
underway. I see people haven't stopped coming to us for samples and more projects. It is
an exciting time. We love what we are doing. It has been stressful at times. [ indiscernible
] will be even more exciting I guess. Okay. This panel is now open for questions. Everybody
getting hungry? One of the issues I raised -- I'm not sure it is a question, but maybe
anybody can comment on this. I think it might be of interest to the audience. Major technology
advances are -- and of course you have heard about this. It's possible to change. It is
not just true for big centers. You could imagine that what is going on now with the sequencing
technologies is that like every six months another instrument becomes available. For
individual investigators, that is a huge commitment for them to purchase one of these instruments.
The are constantly wondering, should we make a major investment in buying 10 of these machines?
There is also just hundreds of individual investigators trying to decide, should I buy
this instrument or wait six months and buy the next one? It is not just the price of
buying the instrument. It is the personnel investment in getting that instrument to work
well in the laboratory. Of course, the computational challenges that Vivian spoke of. Every instrument,
the output is different. It has different nuances. It is a blessing in a curse. Technologies
are revolutionary in their nature but they come with it some hard decisions, especially
when money is tight. That's a lot of money to raise if it is going to be an outdated
machine in the year. There is a lot of nuances associated with that that shouldn't be underestimated.
I don't know if anyone wants to comment, but I was just making that observation. I guess
it is a question of which problem you want to have. Right, right. It is so different
than what the situation was for a long time. Throughout the project really for a few years
beyond, we all couldn't believe we were still doing the classic germination sequencing.
We couldn't believe we were dealing with basically one company, maybe two. At least everything
was very standard and very stable and we got the price down incrementally, but then you
have to be careful what you wish for because now you have the other extreme where there
are so many different things that you can't figure out where to put your money down. Sort
of related question. I am wondering if NIH and, in fact, the federal government is making
plans for resources to be able to analyze all this data. Having myself being tested
and having thousands of genomic traits imitated, I have no idea what to do with that because
analytic tools don't really exist. Is there another phase here that is being anticipated
and planned for? Absolutely. I was going to make some comments. Vivian set it up a little.
I was going to make some comment about that at the very end. Without a question, data
production is not limited now. Computational biology is becoming just the traditional challenges
becoming a bottleneck. It is being discussed extensively at NIH. It is good -- it is getting
considerable amount of discussion. Not only at NIH but the whole industry growing up around
it. [ indiscernible ] in Boston this past week. Venture capitalists are interested.
A lot of different kinds of companies with different models. Actually, it will go hand
in hand. The collection of the data with human genome sequences with phenotype data and medical
records Will Drive the research to add meaning to the individuals to the interpretation but
also how to represent that to consumer independent of the medical profession and how to represent
it to the medical profession so that they cannot have to be absolute experts in every
condition in all the inns and outs of human genetics, but could trick these data much
the same way they treat many of the other biochemical tests you have when you go into
a clinic and be able to sort out, what do I do with this information with this patient?
If I heard you're right, your question was, what is NIH doing? Well, are there resources?
In other words, is there planning? Let me talk to you a little bit about that, and to
lead off from what Eric was saying. A lot of institutes [ indiscernible ] institutionally,
but we have seen this issue across so many different institutes. In fact, when I did
my meeting, we had 15 people from different institutes coming in because I knew they had
the same problem. One of the things that I am taking a hold of is looking at what does
the cloud provide for us to do this kind of analysis? In conjunction with that, the Center
for information technology, which is basically the nuts and bolts of the computing, I have
been in discussions with them as well. They provide the nuts and bolts for us to do that.
Working with different institutes and coming together to educate on how we can match our
forces is one of the discussions. The second one is working with IT to figure out how we
do it. For example, they recognize there may be a need for having a cloud within its context.
The reason for that is the volume of data in the security issue as well. The third piece
to that puzzle is that just started exploring some work with third parties, both commercial
and epidemic in terms of how we do that. They are very interested in working with us. Those
are the three areas that couple what NIH is doing. Just to extend what Jeff said, when
you have your cloud computing meeting, Jeff was saying that private industry [ indiscernible
] but with some of the participants? Google was very interested and Amazon and Microsoft.
They all showed up, right? Yep. 10 years ago they would yawn and say, not interested. Now
they are coming. The other thing is we have a national labs. They do a lot of computing.
They are very interested in this as well. I think we've said this before where we are
now becoming interesting to these communities, because that data wasn't statistically large
or complicated enough. I was happy to see on your slide that you are mentioning environmental
scanning and other areas that need to be integrated. Absolutely. I think we need to not be [ indiscernible
]. One more question before lunch. I am always Rosenbaum. Just to get back to your comment
about the individual sequencing centers facing these challenges. What sort of efforts are
there so that perhaps there is more consolidation? I know that [ indiscernible ] offers sequencing
from other institutions, but it is challenging scheduling. How might that be able to be arranged
so that could be more productive and less effort for individual sites? Anybody want
to take that? [ indiscernible ] I think are organizing that, but is that limiting what
options there are in the future so this could be done more collectively? Is Adam still here?
Jeff, do you want to -- Well, there is a lot of debate whether consolidation into a few
large centers is the best way to go versus trying to disseminate. We are supporting both
approaches. A lot of people don't want to have to wait for large centers to be able
to do the production. Actually, you can also -- most of -- many of the companies, many
sequencing companies and other companies, run sequencing services. If you all want to
step a sequence, you can just send your sample and get it back in a matter of a few weeks.
All the models are being supported. We don't want to sort of converge on a single model,
because we think that would be too constraining. Part of the value of these instruments is
that at some level, you don't have to have as large an infrastructure to do a significant
sequencing project as you had to even three or four years ago, but there are some costs.
Another thing to add there is I had your question about how we work with these other institutes.
Many projects, a lot of these older sequencing centers -- [ indiscernible ] they are all
working on these projects. Our experience is that they have to work together more closely
because when you are generating base pairs from a particular project, it is not related
to a chromosome. All the base pairs are going to a part. We have to analyze it. That is
one thing. The other thing is I have an example with Washington University St. Louis. They
are getting a lot of people coming to them saying, I want to work on smaller projects.
They are working with these guys. They are sharing their knowledge and telling us a little
about it. I think as Jeff said, each one of the different institutes is doing this as
well. It happens on couple levels. Okay. I want to think this panel for their talk.
[ applause ] It is lunchtime, but we need to keep moving on the schedule. Right outside
these doors are some box lunches. You are all welcome to grab lunch, bring it back to
your place. We are going to start up in about 10 minutes. While you are eating, our lunch
speaker will be sharing.