WEBVTT

00:00:00.000 --> 00:00:05.450 align:middle line:90%


00:00:05.450 --> 00:00:07.160 align:middle line:90%
So, this is Julian Parkhill.

00:00:07.160 --> 00:00:11.000 align:middle line:84%
He's going to be telling us
about genomic epidemiology.

00:00:11.000 --> 00:00:14.450 align:middle line:84%
So, Julian, what is
genomic epidemiology?

00:00:14.450 --> 00:00:16.590 align:middle line:84%
Genomic epidemiology
is the process

00:00:16.590 --> 00:00:20.330 align:middle line:84%
of using whole genome sequences
to understand transmission

00:00:20.330 --> 00:00:21.800 align:middle line:90%
of bacterial pathogens.

00:00:21.800 --> 00:00:24.050 align:middle line:84%
So, that can be transmission
within hospitals

00:00:24.050 --> 00:00:26.910 align:middle line:84%
or it can be food-borne
outbreaks, for example.

00:00:26.910 --> 00:00:30.280 align:middle line:84%
So is genomics a
type of genotyping?

00:00:30.280 --> 00:00:33.280 align:middle line:84%
So, genomics is the
ultimate genotyping.

00:00:33.280 --> 00:00:36.530 align:middle line:84%
It's genotyping with
full resolution,

00:00:36.530 --> 00:00:37.880 align:middle line:90%
all the resolution you can get.

00:00:37.880 --> 00:00:41.974 align:middle line:84%
So, if organisms are
identical at the genome level,

00:00:41.974 --> 00:00:43.640 align:middle line:84%
if they have identical
genome sequences,

00:00:43.640 --> 00:00:45.590 align:middle line:90%
then they are identical.

00:00:45.590 --> 00:00:49.460 align:middle line:84%
There is no further resolution
you can do to separate them.

00:00:49.460 --> 00:00:52.070 align:middle line:84%
But that means that you
have a much, much finer

00:00:52.070 --> 00:00:54.370 align:middle line:84%
resolution than you do
with other genotyping.

00:00:54.370 --> 00:00:56.360 align:middle line:84%
And a finer resolution
means that you

00:00:56.360 --> 00:01:00.230 align:middle line:84%
get a better understanding
of transmission,

00:01:00.230 --> 00:01:01.940 align:middle line:84%
you can get a
finer understanding

00:01:01.940 --> 00:01:04.849 align:middle line:84%
of dating and times
of transmission.

00:01:04.849 --> 00:01:08.150 align:middle line:84%
So are there any other
advantages of whole genome

00:01:08.150 --> 00:01:10.910 align:middle line:84%
sequencing over the
more traditional methods

00:01:10.910 --> 00:01:13.160 align:middle line:90%
of molecular typing?

00:01:13.160 --> 00:01:16.860 align:middle line:84%
So, the main advantage is
resolution and accuracy.

00:01:16.860 --> 00:01:18.830 align:middle line:84%
But the other is
interoperability.

00:01:18.830 --> 00:01:22.670 align:middle line:84%
So, the ability of people
around the world to know

00:01:22.670 --> 00:01:25.440 align:middle line:84%
that you're working with
exactly the same thing.

00:01:25.440 --> 00:01:29.780 align:middle line:84%
And with a lot of molecular
techniques like PFGE,

00:01:29.780 --> 00:01:33.650 align:middle line:84%
it's impossible to be
certain that the type you're

00:01:33.650 --> 00:01:36.380 align:middle line:84%
looking at in America is exactly
the same as the type you're

00:01:36.380 --> 00:01:39.380 align:middle line:90%
looking at in the UK.

00:01:39.380 --> 00:01:41.810 align:middle line:84%
With whole genome sequencing,
the data is digital,

00:01:41.810 --> 00:01:44.090 align:middle line:90%
the data is exchangeable.

00:01:44.090 --> 00:01:45.650 align:middle line:84%
And you can be
absolutely certain

00:01:45.650 --> 00:01:47.840 align:middle line:84%
that the strain you're
looking at in the UK

00:01:47.840 --> 00:01:50.120 align:middle line:84%
is exactly the
same as the isolate

00:01:50.120 --> 00:01:51.770 align:middle line:84%
that you're looking
at in America.

00:01:51.770 --> 00:01:55.590 align:middle line:84%
So, I understand one of
the advantages of using

00:01:55.590 --> 00:01:58.730 align:middle line:84%
the traditional
genotyping schemes

00:01:58.730 --> 00:02:02.120 align:middle line:84%
for molecular epidemiology
is that we have established

00:02:02.120 --> 00:02:05.580 align:middle line:84%
databases and that
different labs can compare

00:02:05.580 --> 00:02:07.430 align:middle line:90%
their results to one another.

00:02:07.430 --> 00:02:09.190 align:middle line:84%
Can you do that
kind of thing when

00:02:09.190 --> 00:02:11.000 align:middle line:90%
you're using genome sequencing?

00:02:11.000 --> 00:02:11.570 align:middle line:90%
Absolutely.

00:02:11.570 --> 00:02:13.852 align:middle line:84%
You can compare results
if you exchange data.

00:02:13.852 --> 00:02:16.310 align:middle line:84%
And one of the good things
about whole genome sequence data

00:02:16.310 --> 00:02:22.400 align:middle line:84%
is it is digital and it
is easily exchangeable,

00:02:22.400 --> 00:02:26.120 align:middle line:84%
so that the data is
easily comparable from one

00:02:26.120 --> 00:02:28.460 align:middle line:90%
site to another.

00:02:28.460 --> 00:02:33.170 align:middle line:84%
The disadvantage, I
suppose, is that you

00:02:33.170 --> 00:02:37.000 align:middle line:84%
don't have easily human
understandable nomenclature.

00:02:37.000 --> 00:02:40.610 align:middle line:84%
You can't say this is an-- well,
you can say, this is an SD 22,

00:02:40.610 --> 00:02:44.270 align:middle line:84%
but obviously you're generating
data that's-- information

00:02:44.270 --> 00:02:49.490 align:middle line:84%
that's much more
detailed than that.

00:02:49.490 --> 00:02:53.520 align:middle line:84%
So, while you can, with
traditional typing schemes,

00:02:53.520 --> 00:02:58.040 align:middle line:84%
use a human accessible handle
that everyone understands:

00:02:58.040 --> 00:03:01.850 align:middle line:84%
this is an SD 22,
this is a PFGE Type 4.

00:03:01.850 --> 00:03:04.555 align:middle line:84%
It's much more difficult
with genome sequences.

00:03:04.555 --> 00:03:06.290 align:middle line:84%
Generally, they're
backwards compatible

00:03:06.290 --> 00:03:12.050 align:middle line:84%
so that you can certainly
say, this is an SD 22.

00:03:12.050 --> 00:03:14.360 align:middle line:84%
People are developing
hierarchical schemes

00:03:14.360 --> 00:03:18.860 align:middle line:84%
that allow you to provide
a universal nomenclature.

00:03:18.860 --> 00:03:22.100 align:middle line:84%
But it is going to be
less human understandable.

00:03:22.100 --> 00:03:29.790 align:middle line:84%
But equally, the ability of the
computers to compare that data

00:03:29.790 --> 00:03:33.710 align:middle line:84%
and to understand and to
interpret relationships

00:03:33.710 --> 00:03:34.760 align:middle line:90%
is much, much finer.

00:03:34.760 --> 00:03:36.880 align:middle line:84%
So, perhaps that
level of nomenclature

00:03:36.880 --> 00:03:39.440 align:middle line:90%
isn't quite so important.

00:03:39.440 --> 00:03:41.660 align:middle line:90%
So, it sounds great.

00:03:41.660 --> 00:03:44.280 align:middle line:90%
Why isn't everyone using it yet?

00:03:44.280 --> 00:03:45.170 align:middle line:90%
And

00:03:45.170 --> 00:03:46.440 align:middle line:90%
A number of reasons.

00:03:46.440 --> 00:03:49.590 align:middle line:90%
It's still relatively expensive.

00:03:49.590 --> 00:03:54.050 align:middle line:84%
It takes time and money to build
the infrastructure to do this.

00:03:54.050 --> 00:03:57.170 align:middle line:84%
You need sequencing
machines, you need machines

00:03:57.170 --> 00:04:00.920 align:middle line:84%
for preparing DNA for
the right quality,

00:04:00.920 --> 00:04:03.950 align:middle line:84%
you need people who are trained
in doing that, you need people

00:04:03.950 --> 00:04:08.570 align:middle line:84%
who are trained in interpreting
and understanding the data.

00:04:08.570 --> 00:04:11.870 align:middle line:84%
All of that can be
done, you can put

00:04:11.870 --> 00:04:13.200 align:middle line:90%
in place the infrastructure.

00:04:13.200 --> 00:04:16.130 align:middle line:90%
You can build the expertise.

00:04:16.130 --> 00:04:19.970 align:middle line:84%
But it is, at the moment,
a little more expensive

00:04:19.970 --> 00:04:22.415 align:middle line:84%
than a lot of current
typing techniques,

00:04:22.415 --> 00:04:24.830 align:middle line:84%
certainly the molecular
techniques like PFGE.

00:04:24.830 --> 00:04:30.060 align:middle line:84%
Although, it's probably
cheaper than MLST.

00:04:30.060 --> 00:04:32.720 align:middle line:84%
But building that
infrastructure,

00:04:32.720 --> 00:04:38.100 align:middle line:84%
training those people
takes time and money.

00:04:38.100 --> 00:04:44.610 align:middle line:84%
And running that will also take
a certain investment over time.

00:04:44.610 --> 00:04:48.260 align:middle line:84%
So, for people to do this,
for people to invest in that,

00:04:48.260 --> 00:04:49.950 align:middle line:84%
for governments to
invest in it, it

00:04:49.950 --> 00:04:54.134 align:middle line:84%
requires a proof that this is
a cost-effective thing to do.

00:04:54.134 --> 00:04:55.800 align:middle line:84%
Especially in hospitals,
we have to show

00:04:55.800 --> 00:04:59.730 align:middle line:84%
that routine sequencing is a
cost-effective intervention.

00:04:59.730 --> 00:05:02.190 align:middle line:84%
And then, I think
it will follow.

00:05:02.190 --> 00:05:05.100 align:middle line:84%
So, is this something that
you think will happen?

00:05:05.100 --> 00:05:06.550 align:middle line:90%
I think it's inevitable.

00:05:06.550 --> 00:05:09.930 align:middle line:84%
I think the advantages
are so numerous.

00:05:09.930 --> 00:05:13.430 align:middle line:90%


00:05:13.430 --> 00:05:16.520 align:middle line:84%
Purely in terms of
epidemiology, you

00:05:16.520 --> 00:05:20.830 align:middle line:84%
can generate more accurate data
faster than current techniques.

00:05:20.830 --> 00:05:23.330 align:middle line:84%
That means intervening
with outbreaks earlier.

00:05:23.330 --> 00:05:26.669 align:middle line:84%
It means crucially
disproving outbreaks earlier,

00:05:26.669 --> 00:05:28.460 align:middle line:84%
so you don't spend a
lot of time and effort

00:05:28.460 --> 00:05:32.150 align:middle line:84%
chasing down outbreaks
that aren't real.

00:05:32.150 --> 00:05:34.010 align:middle line:84%
When you add in all the
additional data that

00:05:34.010 --> 00:05:36.050 align:middle line:84%
will come longer
term, drug resistance

00:05:36.050 --> 00:05:39.200 align:middle line:84%
prediction,
virulence prediction,

00:05:39.200 --> 00:05:41.510 align:middle line:90%
it becomes quite compelling.

00:05:41.510 --> 00:05:43.880 align:middle line:84%
And when you think
of one of the things

00:05:43.880 --> 00:05:46.250 align:middle line:84%
that we have to do to
address our current problems

00:05:46.250 --> 00:05:50.720 align:middle line:84%
with antibiotic resistance,
is rational antibiotic use

00:05:50.720 --> 00:05:52.580 align:middle line:90%
and antibiotic stewardship.

00:05:52.580 --> 00:05:57.950 align:middle line:84%
And that means treating
an infection with a drug

00:05:57.950 --> 00:06:02.120 align:middle line:84%
that you know is going to be
effective at that intervention.

00:06:02.120 --> 00:06:05.540 align:middle line:84%
And that means knowing what
somebody is infected with

00:06:05.540 --> 00:06:07.650 align:middle line:84%
and knowing what
it's resistant to.

00:06:07.650 --> 00:06:09.714 align:middle line:84%
And whole genome
sequencing is going

00:06:09.714 --> 00:06:10.880 align:middle line:90%
to be a route to doing that.

00:06:10.880 --> 00:06:13.250 align:middle line:84%
So, I think it's inevitable
that genome sequencing is

00:06:13.250 --> 00:06:15.550 align:middle line:90%
going to be used in this way.

00:06:15.550 --> 00:06:18.880 align:middle line:90%