0:00:14.768,0:00:16.101
Every year,

0:00:16.101,0:00:20.631
700,000 people are killed by superbugs

0:00:21.801,0:00:23.978
To put that in perspective,

0:00:23.978,0:00:30.064
that's more than the combined populations[br]of the cities of Bristol and Bath.

0:00:31.124,0:00:34.812
What's more is that in about thirty years

0:00:34.812,0:00:40.250
that number is expected to rise[br]to 10 million people a year,

0:00:40.250,0:00:45.457
which would mean that superbugs[br]would have killed more people than cancer.

0:00:47.697,0:00:50.137
Now, many of us may have heard [br]the term 'superbug'

0:00:50.137,0:00:52.891
thrown around at some point in our lives.

0:00:52.891,0:00:57.751
But lets take some time to think about[br]what it is we're really referring to,

0:00:57.751,0:00:59.341
how we got here,

0:00:59.341,0:01:00.946
and most importantly,

0:01:00.946,0:01:04.766
how do we get ourselves[br]out of this crisis?

0:01:05.876,0:01:07.704
So what is a superbug?

0:01:08.544,0:01:12.451
'Superbug' is a name given[br]to a particular group of bacteria,

0:01:12.451,0:01:14.607
otherwise known as a strain,

0:01:14.607,0:01:18.426
that cannot be treated[br]with most of the antibiotics

0:01:18.426,0:01:21.223
available or in use today.

0:01:22.113,0:01:25.205
They're a formidable enemy[br]when you consider the fact

0:01:25.205,0:01:29.795
that many people would not have lived[br]to see their sixtieth birthday

0:01:29.795,0:01:34.555
prior to the discovery of penicillin[br]and other classes of antibiotics.

0:01:37.685,0:01:41.354
Raise your hand if you've ever had[br]an antibiotic prescribed to you.

0:01:43.424,0:01:44.514
OK.

0:01:44.514,0:01:48.294
Well, you'll be familiar with the process[br]I'm about to describe next.

0:01:48.644,0:01:52.285
It probably all started[br]when you weren't feeling very well:

0:01:52.285,0:01:54.980
a persistent cough, a fever,

0:01:54.980,0:01:58.841
or perhaps that burning sensation[br]when you'd go for a wee.

0:01:59.931,0:02:02.042
(Laughter)

0:02:03.562,0:02:05.799
Having put up with it all weekend,

0:02:05.799,0:02:10.379
you finally mustered up that courage[br]to go out and speak to your doctor.

0:02:11.229,0:02:13.377
After reviewing your symptoms,

0:02:13.377,0:02:16.633
you're given a prescription[br]for antibiotics and told:

0:02:16.633,0:02:20.577
'Come back after the third dose[br]if your symptoms persist'.

0:02:22.757,0:02:25.719
Now, I'm an engineer,[br]so I love my flowcharts.

0:02:26.659,0:02:30.533
And they really help me[br]get my head around what's going on.

0:02:30.533,0:02:33.752
So I took the liberty of putting [br]this little schematic together.

0:02:34.612,0:02:38.065
So step one: bacterial infection starts.

0:02:38.365,0:02:42.236
Step two: signs and symptoms develop.

0:02:42.946,0:02:48.436
Step three: person goes out [br]to get medical attention.

0:02:49.786,0:02:53.072
Step four: antibiotic[br]prescription is given.

0:02:53.492,0:02:56.809
And step five: after three doses,

0:02:57.799,0:03:00.419
check whether symptoms are getting better.

0:03:00.419,0:03:02.568
If yes, carry on.

0:03:03.338,0:03:09.653
If no, order a diagnostic test [br]to determine what is making you sick,

0:03:09.973,0:03:11.973
and then return to step three.

0:03:13.943,0:03:20.147
Now, I've got the utmost respect[br]for our friends in the medical profession,

0:03:20.147,0:03:24.030
but this system is fundamentally flawed,[br]for a couple of reasons.

0:03:24.490,0:03:28.997
Firstly, it encourages your doctor[br]to start treatment

0:03:28.997,0:03:31.470
before having all the information.

0:03:32.900,0:03:34.431
Secondly,

0:03:34.431,0:03:36.891
each time we go around that loop,

0:03:36.891,0:03:40.224
what we're actually doing[br]is killing off all the bacteria

0:03:40.224,0:03:44.262
that cannot defend themselves[br]against a given antibiotic,

0:03:44.262,0:03:48.649
leaving behind a monolithic group[br]of bacteria that can.

0:03:49.379,0:03:55.191
In effect, we're helping the enemy[br]sift out unfit soldiers.

0:03:58.661,0:04:04.446
Now, we're in this predicament[br]because, still today,

0:04:04.446,0:04:08.089
one of the most commonly used methods[br]for identifying bacteria

0:04:08.089,0:04:13.772
involves taking a sample of urine,[br]of mucus, of blood,

0:04:13.772,0:04:19.315
and growing it under varying conditions[br]to help us piece together the identity.

0:04:19.315,0:04:23.416
Think of it like a process of elimination[br]carried out in the lab.

0:04:24.166,0:04:26.649
Once the bacteria is identified,

0:04:27.119,0:04:28.969
we then use another process

0:04:28.969,0:04:32.906
to determine which antibiotic[br]most likely will work.

0:04:33.566,0:04:36.533
The problem is that that takes time,

0:04:37.043,0:04:38.901
two days or more,

0:04:39.371,0:04:41.735
and in some extreme cases,

0:04:41.735,0:04:46.232
people have actually died[br]before their doctors got the answers.

0:04:48.832,0:04:55.163
Nevertheless, this is our system;[br]treatment first, diagnostic second.

0:04:57.713,0:05:00.007
So how do we get ourselves out of this?

0:05:00.907,0:05:03.256
Actually, it would be a lot better

0:05:03.256,0:05:06.546
if we could first work out[br]what's making you sick,

0:05:06.546,0:05:09.812
and then selecting[br]the most appropriate antibiotic.

0:05:10.352,0:05:16.389
Not only would you get better days sooner,[br]but we'd also curb the rise of superbugs.

0:05:17.509,0:05:21.118
But to do that, [br]we'd need a test that's fast.

0:05:21.118,0:05:24.219
I mean, like, really, really fast;

0:05:24.559,0:05:27.580
fast enough to meet the 20 minutes or less

0:05:27.580,0:05:30.455
that you have with your doctor[br]or your pharmacist.

0:05:31.811,0:05:35.916
But also, it'd also need to be affordable;

0:05:35.916,0:05:38.234
affordable enough[br]that developing countries

0:05:38.234,0:05:39.891
could make that transition.

0:05:41.041,0:05:44.652
And finally, it would need to be[br]easy enough to use

0:05:44.652,0:05:48.481
that it's as effortless[br]as checking your temperature.

0:05:49.191,0:05:53.274
And that is exactly what my team and I[br]have been working towards

0:05:53.274,0:05:55.082
over the last two years.

0:05:56.422,0:06:00.969
We've been developing a technique[br]based on receptor-mediated sensing.

0:06:00.969,0:06:06.443
We've quoted it GMS for short,[br]and it's a lot simpler than it sounds.

0:06:07.373,0:06:10.821
For many bacteria,[br]the first step of infecting you

0:06:10.821,0:06:14.221
involves sticking themselves[br]to the cells in your body

0:06:14.221,0:06:16.594
using a hook and loop system

0:06:16.594,0:06:20.900
that's quite similar[br]to what we've come to love in Velcro.

0:06:20.900,0:06:22.863
(Laughter)

0:06:24.673,0:06:29.496
The hooks in this case[br]would be specific proteins, or receptors,

0:06:29.496,0:06:31.909
on the surface of the bacteria,

0:06:31.909,0:06:36.478
and the loops would be specific molecules[br]on the surface of the cells in your body.

0:06:36.972,0:06:41.682
What we've done is created a low-cost,[br]light-emitting material

0:06:42.132,0:06:44.097
and coated it in loops;

0:06:44.097,0:06:45.587
let's call them probes.

0:06:46.847,0:06:51.254
When we mix our probes with a sample -[br]for example, urine -

0:06:51.254,0:06:53.881
they stick to bacteria like Velcro.

0:06:54.311,0:06:56.513
And by measuring the light [br]coming off of them,

0:06:56.513,0:07:00.736
we can then calculate the number[br]of bacteria present in that sample.

0:07:01.386,0:07:04.552
The final process itself[br]will be quite simple.

0:07:04.552,0:07:08.226
Take a bit of urine, [br]add it to a special cartridge.

0:07:08.226,0:07:12.191
Then place that cartridge[br]in our very own bug detector.

0:07:14.139,0:07:17.387
The cartridge would mix[br]the urine with our probes

0:07:17.387,0:07:22.035
before separating out the bacteria[br]and measuring the light coming off.

0:07:22.035,0:07:26.441
That final step will allow us to determine[br]whether bacteria is present,

0:07:26.441,0:07:29.345
which one, and how many,

0:07:29.345,0:07:31.770
all within 15 minutes.

0:07:32.185,0:07:35.613
From two days, to 15 minutes.

0:07:37.443,0:07:40.431
Now, as with most scientific developments,

0:07:40.431,0:07:45.709
the journey is fraught with challenges[br]and sometimes disappointments, frankly.

0:07:47.199,0:07:49.426
And I'll share a couple of them with you.

0:07:50.086,0:07:53.654
One of our first challenges was: [br]how do you develop a probe

0:07:53.654,0:07:56.712
that effectively mimics[br]the cells in your body?

0:07:58.760,0:08:01.841
It took a team of researchers at the [br]University of Bristol

0:08:01.841,0:08:03.853
months of experimenting

0:08:03.853,0:08:06.193
to get the recipe just right

0:08:06.193,0:08:11.022
to have the right type and balance[br]of loops to mimic the cell.

0:08:11.022,0:08:14.620
I mean, this is not like bashing out [br]bad pancakes on a Sunday morning.

0:08:14.620,0:08:17.931
This is the kind of effort[br]you put into winning Masterchef.

0:08:17.931,0:08:20.154
(Chuckling)

0:08:21.074,0:08:25.858
The second challenge for us was actually[br]finding a detection method

0:08:25.858,0:08:30.872
that was low-cost, yet powerful enough[br]to detect the probes

0:08:30.872,0:08:32.874
once attached to the bacteria.

0:08:33.874,0:08:38.098
For that, we turned to a well-established[br]technique based on fluorescence.

0:08:38.903,0:08:40.634
Fluorescence is a phenomenon

0:08:40.634,0:08:44.434
where a material stimulated with light[br]absorbs some of it,

0:08:44.434,0:08:46.446
and re-emits a different colour.

0:08:46.926,0:08:48.896
As a detection technique,

0:08:48.896,0:08:52.186
we'd effectively take[br]a specific colour of light,

0:08:52.186,0:08:55.901
shine it at that material, [br]and observe the colour coming back,

0:08:55.901,0:08:58.400
and that helps us tell what's present.

0:08:59.690,0:09:03.759
So, we've got probes, [br]they stick to the right bacteria,

0:09:04.209,0:09:05.570
and we can detect them.

0:09:05.570,0:09:06.990
Job done, right?

0:09:07.540,0:09:09.124
Actually, not quite.

0:09:10.684,0:09:13.372
Most of the foods that we eat [br]actually get broken down

0:09:13.372,0:09:16.678
into light-emitting materials[br]that end up in your wee.

0:09:17.928,0:09:22.845
And so, these are things like [br]energy drinks, vitamin supplements,

0:09:22.845,0:09:24.389
pregnancy supplements,

0:09:24.389,0:09:28.283
and all of these together[br]could lead to false positives.

0:09:29.113,0:09:33.389
In other words, they make it really hard[br]for us to determine the difference

0:09:33.389,0:09:35.619
between friend or foe.

0:09:36.469,0:09:42.570
And so, differentiating our probes from[br]the remnants of your last energy drink

0:09:42.570,0:09:44.767
is absolutely important

0:09:44.767,0:09:47.847
for ensuring that we do not report[br]that bacteria is present

0:09:47.847,0:09:49.601
even when there aren't any.

0:09:51.111,0:09:54.390
This is where the research[br]took a really interesting turn,

0:09:54.390,0:09:58.557
because getting around this[br]meant getting familiar with our probes,

0:09:58.557,0:10:01.297
but also getting[br]really familiar with urine.

0:10:03.547,0:10:07.852
We carried out several experiments [br]observing how our probes behaved

0:10:07.852,0:10:10.501
when stimulated with[br]different colours of light.

0:10:10.501,0:10:14.283
Particularly, we're interested[br]in how much green light is emitted

0:10:14.283,0:10:17.779
for every colour of light[br]that we stimulated them with.

0:10:18.442,0:10:21.140
This is the profile of our probes,

0:10:21.647,0:10:23.187
and it does not change.

0:10:24.247,0:10:27.677
So any deviation from that[br]would then tell us

0:10:27.677,0:10:30.657
that there's other stuff[br]mixed in with the sample,

0:10:30.657,0:10:35.584
and it would look a little bit like[br]when it's mixed with urine, for example.

0:10:35.894,0:10:42.251
By measuring that change, [br]we can actively correct for interference

0:10:42.251,0:10:45.397
from anything else[br]that's present in the urine.

0:10:46.365,0:10:51.126
So pulling it all together, [br]the probes help us find the bacteria

0:10:52.126,0:10:54.528
and fluorescence helps us find the probes.

0:10:56.608,0:10:59.535
We're still at the[br]early stages of our journey,

0:10:59.535,0:11:03.176
but we were able to take[br]our first prototype into a hospital lab

0:11:03.176,0:11:06.224
where we tested for E. coli [br]in human urine.

0:11:06.944,0:11:11.919
GMS was able to detect the presence [br]of urine without any false positives,

0:11:12.229,0:11:18.504
and it correctly reported the absence [br]of E. coli in the urine in 63% of cases.

0:11:19.494,0:11:20.747
Finally,

0:11:21.107,0:11:23.347
by placing three measurements in parallel,

0:11:23.347,0:11:27.704
we were able to achieve an average time[br]per test of just four minutes.

0:11:30.044,0:11:32.979
This is certainly not[br]the end of the story for us.

0:11:33.489,0:11:37.804
We're now developing other forms of loops[br]to address other harmful bacteria,

0:11:38.214,0:11:42.212
and we aim to test and improve[br]the system a few more times

0:11:42.212,0:11:43.967
before it gets to your GP.

0:11:44.297,0:11:46.250
But once we do this,

0:11:46.637,0:11:50.297
combining our faster, accurate diagnostics

0:11:50.867,0:11:56.670
with better antibiotic usage[br]and continued public engagement,

0:11:56.670,0:12:00.182
we could finally tip this war[br]back in our favour

0:12:00.182,0:12:04.457
and overcome superbugs once and for all.

0:12:04.877,0:12:06.019
Thank you.

0:12:06.019,0:12:07.771
(Applause)