September 24, 2024 • 19 mins
Scientific advances have led us to diagnosing, treating and preventing cancer much better than ever before. Still, 10 million people die of cancer every year and what we need now are transformative ideas. In this episode you’ll hear from Dr Amanda Khoury and Dr Braydon Meyer who are using AI to develop the next generation of cancer treatments and diagnostics.
Support Garvan research: https://www.garvan.org.au/support-us/give-now
See omnystudio.com/listener for privacy information.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Dr Viviane Richter (00:00):
Cancer research has come a long way. Scientific advances have
led us to diagnosing, treating and preventing cancer much better
than ever before. Today, around 70% of people diagnosed with
cancer will survive five years or more, an increase from
around 50% in the late 1980s. Still, 10 million people
(00:23):
die of cancer every year. What we need now are
transformative ideas, and artificial intelligence is helping us get there.
Today you'll hear from two Garvan researchers who are using
AI to develop the next generation of cancer treatments and diagnostics.
You're listening to Medical Minds. The podcast that takes you
(00:44):
inside the labs at the Garvan Institute of Medical Research.
I'm your host, Dr Viviane Richter. And with me here
is Dr Amanda Khoury and Dr Braydon Meyer, researchers in
the Cancer Epigenetics Lab at Garvan. Welcome, Amanda and Braydon.
Dr Amanda Khoury (01:00):
Hi there, Viv. Thanks for having us.
Dr Braydon Meyer (01:01):
Great to be here, Viv. Thank you.
Dr Viviane Richter (01:04):
Amanda and Braydon. Before we get to how you're using
AI in your cancer research, can you tell us what
epigenetics means?
Dr Amanda Khoury (01:13):
So every single human begins as a single cell, and
that single cell contains a copy of your DNA – the complete
set – one half from your mum, one half from your dad.
That cell divides over and over and over again. It
keeps going until you have a whole human. But if
you look at yourself, you've got different types of cells.
You've got eye cells, you've got heart cells, you've got
brain cells and they all look different. Yet, they've got
(01:36):
the same instruction manual within them. So the reason that
we can have these different cell types is because there's
another layer of regulation. It sits above the genome, and it
pulls the strings to tell certain genes in certain cells
to turn on and certain to turn off. And it's
that different profile of expression that gives rise to different
cell types. And that's what epigenetics is.
Dr Viviane Richter (01:56):
So your epigenetics gives rise to the different cell types,
different proteins, etc. What does that mean in the context
of cancer?
Dr Amanda Khoury (02:06):
For a long time, there was the goal to sequence
the human genome because we thought if we crack that sequence,
we're going to understand how the body works. We are
going to be able to find all the mutations that
happen within cancer because we know cancer is a disease
caused by mutations. But, it's not that simple. There's actually
more than that that needs to be understood. And so
the epigenome allows healthy gene expression. So if you have
(02:29):
a healthy epigenome, you're healthy. If you don't, you're not healthy.
So in cancer, what we see is that the epigenome
becomes completely disrupted.
Dr Braydon Meyer (02:38):
As Amanda was saying, we initially thought cancer was just
purely mutations, and these mutations give rise to cancer. And
while that is still very much true, we're kind of
understanding now that it's actually this entanglement of the genome.
So these mutations give rise to the cancer, but it's
the epigenome that modulate it. It could give it more
resistance to therapies. It could make it more malignant or
(03:01):
metastatic or something like that. So, essentially it's the epigenome
and the genome kind of working in tandem. And that's
what gives people somewhat unique cancers, essentially because you can
have unique profiles that might give rise to these different
forms of cancer.
Dr Viviane Richter (03:17):
So the epigenetics is involved in these different forms of cancers.
Can you give us a picture of what the research
question that you're hoping to answer is?
Dr Amanda Khoury (03:28):
As we were saying, the epigenome is multifaceted. It's got
a lot of features, so studying it means really drilling
in and taking your specialty. And my specialty is on
the three dimensional structure of DNA. So we know that
in every cell there's two metres of DNA, and that
means it needs to be incredibly compacted to fit. And
what my research and others has shown is that, rather
(03:50):
than just being sort of screwed up into a ball, it's got
a very precise structure that's really important in telling which
portions to be switched on and which portions to be
switched off. So I study how that structure becomes undone
in cancer.
Dr Viviane Richter (04:04):
What has the 3D structure shown you?
Dr Amanda Khoury (04:07):
So I actually work on the parts of the genome
that are involved in creating those pieces of the structure
coming together. So sort of like the folding points. And what
I found that in cancer you actually get mutations at
these folding points, which is what leads to the disruption
of the genome.
Dr Viviane Richter (04:25):
And how does that relate to cancers?
Dr Amanda Khoury (04:27):
We've looked across a whole range of cancers and we've
shown that, no matter what cancer we look at, you
actually see these points being mutated across the whole set. So,
it sort of points towards this idea of the 3D
genome becoming unwound as a common mechanism that could occur
across cancer and therefore something that could potentially be targeted.
Dr Viviane Richter (04:48):
So how would you target these regions?
Dr Amanda Khoury (04:50):
So that's a very difficult question that is definitely going
to take a lot more research. But essentially, the way
I like to break down cancer research is that you
have to first work out how the system is working
in normal cells. You then have to understand how it's
broken in cancer cells. And then after that, once you
figure out how it's broken, can you actually exploit it?
(05:10):
Can you target it? Some things you can't touch or
undo in the genome. So right now we're at that
point where we discovered this really interesting pattern and we're
sort of working out "can we change it?" And in the
lab we are sort of doing research at the moment,
which is testing drugs that can try to revert the
changes to the 3D genome in order to make it
more 'healthy cell' like. So, while my particular research is a bit far off,
(05:32):
it's pointing in that sort of direction.
Dr Viviane Richter (05:35):
So we know that both of you are working with AI.
Is AI working to speed up that process?
Dr Amanda Khoury (05:41):
Now, the way that I use AI, my research is focusing
on a part of the genome that's really not as
well studied because it's just more mysterious. So you can
break the genome into two parts. You've got the one
that has more obvious function that makes proteins, and you
have the other part, which makes things that we don't
really fully understand what they do. They are vastly different
(06:03):
in size, so the part that we can understand more
clearly is only 2% out of the genome. And that
means that that's only 600 million of the bases, the
600 million of those ATGC, and then 2.4 billion are in
the non-coding section, which we don't understand that well. So
I'm getting to use AI because of this very large data
(06:24):
set that I'm trying to parse to point me in
the right direction of which of these 2.4 billion bases
I should be focusing my research on. So I think
what was really remarkable in what we found was that, okay:
we used AI to help us find these points of interest.
And once we could generate that list of locations we
should be interested in, we could ask lots of research
questions about it. So, for example, I simply asked
(06:47):
seem to be very important in the genome. Are they
mutated in cancer?" I could then take thousands of patient
data sets, overlay them with my list of genomic regions.
And lo and behold, they were very heavily enriched for
being mutated at these sites. Which means that this could
be pointing towards a common pathway that's disrupted in cancer
(07:09):
that involves this unfolding of the 3D genome. And, down the track you know, potentially,
this means that we can maybe find a pathway or
a mechanism that you can target and treat across all cancers.
Dr Viviane Richter (07:22):
Now, Braydon, you're also using AI in your research. Can you
tell us about that?
Dr Braydon Meyer (07:27):
In my research, I'm looking at DNA methylation, which is
an epigenetic mark that we were talking about earlier. And essentially,
instead of how Amanda uses like the the whole 3D
structure of the genome, mine's a little bit more granular. You
kind of go down to the individual bases and you
see these little marks that we call DNA methylation. And
these are known to kind of like switch genes on
(07:49):
and off. But we can also use them just as
a marker. And essentially, in cancer, we can see that
these give each individual cancer – or each type of cancer
rather – their own fingerprint, essentially, which makes them a little
bit unique. But it also allows us to differentiate them
from other tumour types.
Dr Viviane Richter (08:07):
You're using epigenetics as a unique fingerprint for cancer. And
is it right that you're studying breast tumours that look
similar but are actually quite different?
Dr Braydon Meyer (08:17):
Yeah, yeah, exactly. So with the tumour that I'm looking
at, it's called the phyllodes tumour, it's quite rare, and the problem
with that rarity is that it often gets misdiagnosed as
more common tumours because they look morphologically really, really similar.
What we're trying to do is use DNA methylation to
kind of pull these tumours apart because they have this
unique fingerprint. And then, what we can do in terms of AI,
(08:40):
is we take this data that we've generated, these points of
DNA methylation, and then we can put them into an AI
model which allows it to kind of learn what sites
are able to differentiate these tumours the best. What we
did is we took nearly a million points of data
and then distilled it down to 10 that are able
to predict the differences between these tumours alone.
Dr Viviane Richter (09:03):
Could you have done this work without AI?
Dr Braydon Meyer (09:06):
Without AI, painstakingly, I could have done this. There's no way I'm
going to run a million individual tests or something like that.
But yeah, I think this goes back to the point
of Amanda being like "this just really speeds up the process".
Dr Viviane Richter (09:19):
Tell us more about how this misdiagnosis happens at the
moment in the clinic.
Dr Braydon Meyer (09:24):
The problem is that the tumours that I'm trying to differentiate,
these all have really, really similar characteristics. They all, like,
mimic one another under the microscope. But the thing that
we know is that, at the molecular level, we can
see the DNA methylation very, very strikingly different between these tumours.
So that's where we can use this, plus AI, to really differentiate
(09:45):
these tumours where a pathologist might struggle.
Dr Viviane Richter (09:49):
Braydon, you're studying phyllodes tumours. How are these tumours currently diagnosed?
Dr Braydon Meyer (09:54):
So these tumours are currently diagnosed by pathologists. They essentially
just look down the microscope. They're looking for some very
specific features that make it obvious that it's this tumour type.
The only problem that they have is that there are
two other tumours. So, on the benign end of the spectrum,
the phyllodes tumour looks like a fibroadenoma. And on the malignant
(10:15):
end of the spectrum, these tumours look like a metaplastic breast cancer. And
this is really problematic for these patients because if they
get misdiagnosed, say for instance, as a metaplastic breast cancer, when they actually
have a malignant phyllodes tumour, then they'll receive a form of
chemotherapy that is going to be completely useless on a
phyllodes tumour. The phyllodes tumours have no treatments available to
(10:37):
them at the moment. So it's really important that they
get the diagnosis correct so that they don't have these
incorrect treatments given to them.
Dr Viviane Richter (10:45):
Your research aims to prevent women from having to go
through this unnecessary treatment.
Dr Braydon Meyer (10:50):
Exactly. Yeah. So we're just trying to stop them from
being treated inappropriately because, on the benign end of the spectrum,
these patients, they'll have a slightly different surgery that occurs.
So we need to make sure that they're getting the
appropriate surgeries, making sure they are not being overtreated. They're
not receiving therapy that they shouldn't be receiving. The great thing
about this work, there's a functional aspect to this, so we can learn what
(11:12):
genes are potentially being disrupted in these specific tumours. We
know very, very little about them, and the more that
we do learn about them, the more we open up
that Pandora's box, where we can potentially lead to the
first therapeutic intervention for these tumours as well.
Dr Viviane Richter (11:28):
So, Braydon, how would this work in the clinic?
Dr Braydon Meyer (11:31):
So essentially, we're hoping we can take this signature that
we've found and develop it into a test utilising technology
that's already available in pathology labs all across the country.
We're kind of hoping that this test would simply be
able to be done in less than a day, and
it would just give this binary answer. It's going to
be either a phyllodes tumour or a fibroadenoma, or a metaplastic breast cancer in this particular case.
Dr Viviane Richter (11:55):
Can you talk to us about how you collaborate on
your research?
Dr Braydon Meyer (11:59):
So Amanda and I, we joke that between the two
of us, we're one functional scientist. Uh, when you drill down a little bit deeper though, we're definitely, like, very individual.
But in a really good way. I am what they
call the dry lab. I'm just doing a lot of
data analysis. I'm the one generating these models, these AI models that we're
talking about. Whereas, Amanda is more in the wet lab
(12:22):
and it's actually the interaction that makes it so much
more powerful because you don't want to be a jack
of all trades and a master of none. It's much,
much better to be an expert in one field and
then pass on your results to somebody else who is
an expert in their field.
Dr Amanda Khoury (12:37):
Yeah, Braydon sits in the office behind me, and he
bugs me throughout the day.
Dr Braydon Meyer (12:41):
All the time.
Dr Amanda Khoury (12:42):
All the time. Yeah. So my job is to come
up with the experimental designs, design the really complicated, you know,
what parts of the genome are we going to target? Can I
sequence some DNA to touch this bit and not that bit?
It's very intensely complex. And then I do those, what
I like to say, perfect experiments. And then I'll pass
it on to somebody like Braydon, who picks it up
(13:02):
and gets to interpret the results.
Dr Braydon Meyer (13:04):
And it's an iterative process as well. Amanda can pass
that data to me. I'll do my analysis, which might
then feed back to Amanda that instead of, okay, we
started with this whole genome analysis or something like that.
And now we're down to, okay, we have, like, three genes that
we need to look at. I give them back to Amanda.
She looks at those three genes experimentally. She tries to
(13:24):
knock them out, tries to figure out what it is
that their function and how they work. And then again,
she can pass that back to me. I'll do the
data analysis on her experiments.
Dr Viviane Richter (13:33):
How did you guys end up at Garvan?
Dr Braydon Meyer (13:35):
So my interest in science from the very beginning has
been epigenetics. I did it in my honours. I did
it in my PhD back in Melbourne at the Murdoch
Children's Research Institute. And then I had an opportunity, essentially, to
to come to Sydney and, yeah, work at the Garvan. This
was very specifically, so I could work in the lab
(13:55):
of Professor Susan Clark and Associate Professor Clare Stirzaker. Just because
those two are at the absolute forefront of epigenetic research
in Australia.
Dr Amanda Khoury (14:05):
Braydon's being a little bit humble here. Um, he actually
got offered a job in Singapore before he was offered
this job, but he turned it down because of how
much he wanted to work in our lab.
Dr Viviane Richter (14:14):
Amanda, where did your love for epigenetics begin?
Dr Amanda Khoury (14:18):
So I'm a bit of a latecomer to science. I,
in school, I studied the humanities, history, language, went to uni,
was doing a liberal arts degree, decided to try a
neuroscience subject. I thought that sounds kind of interesting. And
then that was where the 'wow' moment happened. I thought, oh,
my goodness. I've never thought about how powerful cells are.
(14:40):
Cells are what make you happy. Cells are what make
you healthy. Cells are what make you think. And I
thought these little powerhouses exist, and I've never thought about
them before. And I just sort of got obsessed with this
concept and then, literally, just did a full pivot into
a medical science degree. I really wanted to work at
Garvan because Garvan is one of the best research institutes
in Australia. And so for me, all I wanted to
(15:03):
do was get my foot in the door. I got
my foot in the door. I was working in this lab,
which was sort of doing routine drug testing for cancer.
It was interesting, but it wasn't stimulating enough. I wanted
to do more thinking about those mechanisms and what was happening.
And then I discovered the Epigenetics Lab, and I thought,
"that's it". And so then I saw Professor Sue Clark
and I said, "Can I please join your lab?" And
(15:24):
that's where the PhD began.
Dr Viviane Richter (15:26):
It's no secret that science can be a tough slog. Amanda,
what motivates you day to day? What makes you keep going?
Dr Amanda Khoury (15:34):
So I often think about the fact that you know,
I wake up, I get out of bed, I go
to work, you know, do my job, come back home.
It all seems really simple, but during the day there
are millions and millions of processes happening automatically in my
body that are allowing me to do that, and I
just find that really wondrous. And the thing is, these
processes are happening. But we're still finding out why they're happening,
(15:55):
what the mechanisms are. We still don't really know how
we tick. So what gets me out of bed is
really trying to answer those fundamental questions about why I exist,
how I exist.
Dr Viviane Richter (16:06):
Braydon, what about you?
Dr Braydon Meyer (16:07):
So, for me personally, I'm fascinated by rare cancers, and
there's quite an inequity in research across the globe for
certain cancers and certain diseases. There is absolutely no work
being done, for instance, on the phyllodes tumour or a
bunch of other rare cancers as well. But just because
these cancers are rare and they affect a very small
(16:28):
proportion of the population, doesn't mean that they shouldn't have
treatment options available for them.
Dr Viviane Richter (16:35):
Amanda, what is a big thing that you would like
to achieve in your career?
Dr Amanda Khoury (16:38):
So I've spoken about this idea that we've sequenced the
human genome, and then we thought that would hold all
the answers. And then we discovered the fact that the
DNA is actually not even linear. It's got a 3D structure.
But that finding only happened maybe 12 years ago. And
so for me, I'm absolutely fascinated by the three dimensional structure.
So I really want my role to be about solidifying
(16:59):
that structure and understanding how it's broken down, not just
in cancer, but in every disease. You know, if we find these
common mechanisms, that we could potentially find a common therapeutic target.
The really important thing about the research that Braydon and
I are doing is that the discoveries we're making, the
methods we're creating, they are not just applicable to single cancer types.
They are applicable to a broad range of cancers. So
(17:22):
what this means is we are doing research smarter. We
are combining AI, we are combining collaborative efforts, in order to really
find the fastest, smartest way to get more therapies to
more patients.
Dr Viviane Richter (17:35):
Amanda and Braydon, before we let you get back to the lab,
it's time for the Fast Five. Amanda, what was your
first job?
Dr Amanda Khoury (17:42):
I had the worst first job ever. I was the
woman in the perfume department handing out the samples.
Dr Braydon Meyer (17:49):
Look at you now.
Dr Viviane Richter (17:50):
Still very fashionable.
Dr Amanda Khoury (17:52):
Oh, thank you.
Dr Viviane Richter (17:54):
Do you have a favourite movie?
Dr Braydon Meyer (17:55):
I'd say my favourite movie would be Across The Universe.
It's a Beatles musical.
Dr Viviane Richter (18:02):
What is the current book you're reading?
Dr Amanda Khoury (18:03):
I'm currently reading two books. One is called Caledonian Road,
which I'm absolutely loving. It is about class inequality and
art and philosophy, and I'm also reading a Gentleman in Moscow,
which was recently ranked by The New York Times as
the third best book ever.
Dr Viviane Richter (18:19):
Something that's on your bucket list?
Dr Braydon Meyer (18:21):
My bucket list is I want to, like, step on
every continent. The major one being Antarctica. I'd love to go
and just hang out with penguins for a bit or
see the northern lights. That's a big one for me too.
Dr Viviane Richter (18:34):
Who inspires you the most?
Dr Amanda Khoury (18:35):
My parents. So they came to Australia to escape the
civil war in Lebanon, and my dad was the eldest
of nine kids. He couldn't go to school because he
had to just start doing labour jobs to support his
whole family. My parents raised a family of six kids
on one income, and we're now all university educated. So it's
a real step up. So they're my inspiration.
Dr Viviane Richter (18:57):
Doctor Amanda Khoury. Doctor Braydon Meyer. Thank you so much
for joining us on Medical Minds.
Dr Amanda Khoury (19:03):
It's been so good being here.
Dr Braydon Meyer (19:04):
It's been great. Thanks for having us.
Dr Viviane Richter (19:06):
If you'd like to know more about this research or
donate to the work we do at Garvan, head to garvan.org.au.
And if you've enjoyed this podcast, please leave a review
and share with other podcast lovers. I'm Dr Viviane Richter.
Thanks for listening.
This podcast was recorded on the traditional Country of the
(19:27):
Gadigal people of the Eora Nation. We recognise their continuing
connection to land, waters and community. We pay our respect
to Aboriginal and Torres Strait Islander cultures and Elders past,
present and emerging.
Cancer research has come a long way. Scientific advances have
led us to diagnosing, treating and preventing cancer much better
than ever before. Today, around 70% of people diagnosed with
cancer will survive five years or more, an increase from
around 50% in the late 1980s. Still, 10 million people
(00:23):
die of cancer every year. What we need now are
transformative ideas, and artificial intelligence is helping us get there.
Today you'll hear from two Garvan researchers who are using
AI to develop the next generation of cancer treatments and diagnostics.
You're listening to Medical Minds. The podcast that takes you
(00:44):
inside the labs at the Garvan Institute of Medical Research.
I'm your host, Dr Viviane Richter. And with me here
is Dr Amanda Khoury and Dr Braydon Meyer, researchers in
the Cancer Epigenetics Lab at Garvan. Welcome, Amanda and Braydon.
Dr Amanda Khoury (01:00):
Hi there, Viv. Thanks for having us.
Dr Braydon Meyer (01:01):
Great to be here, Viv. Thank you.
Dr Viviane Richter (01:04):
Amanda and Braydon. Before we get to how you're using
AI in your cancer research, can you tell us what
epigenetics means?
Dr Amanda Khoury (01:13):
So every single human begins as a single cell, and
that single cell contains a copy of your DNA – the complete
set – one half from your mum, one half from your dad.
That cell divides over and over and over again. It
keeps going until you have a whole human. But if
you look at yourself, you've got different types of cells.
You've got eye cells, you've got heart cells, you've got
brain cells and they all look different. Yet, they've got
(01:36):
the same instruction manual within them. So the reason that
we can have these different cell types is because there's
another layer of regulation. It sits above the genome, and it
pulls the strings to tell certain genes in certain cells
to turn on and certain to turn off. And it's
that different profile of expression that gives rise to different
cell types. And that's what epigenetics is.
Dr Viviane Richter (01:56):
So your epigenetics gives rise to the different cell types,
different proteins, etc. What does that mean in the context
of cancer?
Dr Amanda Khoury (02:06):
For a long time, there was the goal to sequence
the human genome because we thought if we crack that sequence,
we're going to understand how the body works. We are
going to be able to find all the mutations that
happen within cancer because we know cancer is a disease
caused by mutations. But, it's not that simple. There's actually
more than that that needs to be understood. And so
the epigenome allows healthy gene expression. So if you have
(02:29):
a healthy epigenome, you're healthy. If you don't, you're not healthy.
So in cancer, what we see is that the epigenome
becomes completely disrupted.
Dr Braydon Meyer (02:38):
As Amanda was saying, we initially thought cancer was just
purely mutations, and these mutations give rise to cancer. And
while that is still very much true, we're kind of
understanding now that it's actually this entanglement of the genome.
So these mutations give rise to the cancer, but it's
the epigenome that modulate it. It could give it more
resistance to therapies. It could make it more malignant or
(03:01):
metastatic or something like that. So, essentially it's the epigenome
and the genome kind of working in tandem. And that's
what gives people somewhat unique cancers, essentially because you can
have unique profiles that might give rise to these different
forms of cancer.
Dr Viviane Richter (03:17):
So the epigenetics is involved in these different forms of cancers.
Can you give us a picture of what the research
question that you're hoping to answer is?
Dr Amanda Khoury (03:28):
As we were saying, the epigenome is multifaceted. It's got
a lot of features, so studying it means really drilling
in and taking your specialty. And my specialty is on
the three dimensional structure of DNA. So we know that
in every cell there's two metres of DNA, and that
means it needs to be incredibly compacted to fit. And
what my research and others has shown is that, rather
(03:50):
than just being sort of screwed up into a ball, it's got
a very precise structure that's really important in telling which
portions to be switched on and which portions to be
switched off. So I study how that structure becomes undone
in cancer.
Dr Viviane Richter (04:04):
What has the 3D structure shown you?
Dr Amanda Khoury (04:07):
So I actually work on the parts of the genome
that are involved in creating those pieces of the structure
coming together. So sort of like the folding points. And what
I found that in cancer you actually get mutations at
these folding points, which is what leads to the disruption
of the genome.
Dr Viviane Richter (04:25):
And how does that relate to cancers?
Dr Amanda Khoury (04:27):
We've looked across a whole range of cancers and we've
shown that, no matter what cancer we look at, you
actually see these points being mutated across the whole set. So,
it sort of points towards this idea of the 3D
genome becoming unwound as a common mechanism that could occur
across cancer and therefore something that could potentially be targeted.
Dr Viviane Richter (04:48):
So how would you target these regions?
Dr Amanda Khoury (04:50):
So that's a very difficult question that is definitely going
to take a lot more research. But essentially, the way
I like to break down cancer research is that you
have to first work out how the system is working
in normal cells. You then have to understand how it's
broken in cancer cells. And then after that, once you
figure out how it's broken, can you actually exploit it?
(05:10):
Can you target it? Some things you can't touch or
undo in the genome. So right now we're at that
point where we discovered this really interesting pattern and we're
sort of working out "can we change it?" And in the
lab we are sort of doing research at the moment,
which is testing drugs that can try to revert the
changes to the 3D genome in order to make it
more 'healthy cell' like. So, while my particular research is a bit far off,
(05:32):
it's pointing in that sort of direction.
Dr Viviane Richter (05:35):
So we know that both of you are working with AI.
Is AI working to speed up that process?
Dr Amanda Khoury (05:41):
Now, the way that I use AI, my research is focusing
on a part of the genome that's really not as
well studied because it's just more mysterious. So you can
break the genome into two parts. You've got the one
that has more obvious function that makes proteins, and you
have the other part, which makes things that we don't
really fully understand what they do. They are vastly different
(06:03):
in size, so the part that we can understand more
clearly is only 2% out of the genome. And that
means that that's only 600 million of the bases, the
600 million of those ATGC, and then 2.4 billion are in
the non-coding section, which we don't understand that well. So
I'm getting to use AI because of this very large data
(06:24):
set that I'm trying to parse to point me in
the right direction of which of these 2.4 billion bases
I should be focusing my research on. So I think
what was really remarkable in what we found was that, okay:
we used AI to help us find these points of interest.
And once we could generate that list of locations we
should be interested in, we could ask lots of research
questions about it. So, for example, I simply asked
(06:47):
seem to be very important in the genome. Are they
mutated in cancer?" I could then take thousands of patient
data sets, overlay them with my list of genomic regions.
And lo and behold, they were very heavily enriched for
being mutated at these sites. Which means that this could
be pointing towards a common pathway that's disrupted in cancer
(07:09):
that involves this unfolding of the 3D genome. And, down the track you know, potentially,
this means that we can maybe find a pathway or
a mechanism that you can target and treat across all cancers.
Dr Viviane Richter (07:22):
Now, Braydon, you're also using AI in your research. Can you
tell us about that?
Dr Braydon Meyer (07:27):
In my research, I'm looking at DNA methylation, which is
an epigenetic mark that we were talking about earlier. And essentially,
instead of how Amanda uses like the the whole 3D
structure of the genome, mine's a little bit more granular. You
kind of go down to the individual bases and you
see these little marks that we call DNA methylation. And
these are known to kind of like switch genes on
(07:49):
and off. But we can also use them just as
a marker. And essentially, in cancer, we can see that
these give each individual cancer – or each type of cancer
rather – their own fingerprint, essentially, which makes them a little
bit unique. But it also allows us to differentiate them
from other tumour types.
Dr Viviane Richter (08:07):
You're using epigenetics as a unique fingerprint for cancer. And
is it right that you're studying breast tumours that look
similar but are actually quite different?
Dr Braydon Meyer (08:17):
Yeah, yeah, exactly. So with the tumour that I'm looking
at, it's called the phyllodes tumour, it's quite rare, and the problem
with that rarity is that it often gets misdiagnosed as
more common tumours because they look morphologically really, really similar.
What we're trying to do is use DNA methylation to
kind of pull these tumours apart because they have this
unique fingerprint. And then, what we can do in terms of AI,
(08:40):
is we take this data that we've generated, these points of
DNA methylation, and then we can put them into an AI
model which allows it to kind of learn what sites
are able to differentiate these tumours the best. What we
did is we took nearly a million points of data
and then distilled it down to 10 that are able
to predict the differences between these tumours alone.
Dr Viviane Richter (09:03):
Could you have done this work without AI?
Dr Braydon Meyer (09:06):
Without AI, painstakingly, I could have done this. There's no way I'm
going to run a million individual tests or something like that.
But yeah, I think this goes back to the point
of Amanda being like "this just really speeds up the process".
Dr Viviane Richter (09:19):
Tell us more about how this misdiagnosis happens at the
moment in the clinic.
Dr Braydon Meyer (09:24):
The problem is that the tumours that I'm trying to differentiate,
these all have really, really similar characteristics. They all, like,
mimic one another under the microscope. But the thing that
we know is that, at the molecular level, we can
see the DNA methylation very, very strikingly different between these tumours.
So that's where we can use this, plus AI, to really differentiate
(09:45):
these tumours where a pathologist might struggle.
Dr Viviane Richter (09:49):
Braydon, you're studying phyllodes tumours. How are these tumours currently diagnosed?
Dr Braydon Meyer (09:54):
So these tumours are currently diagnosed by pathologists. They essentially
just look down the microscope. They're looking for some very
specific features that make it obvious that it's this tumour type.
The only problem that they have is that there are
two other tumours. So, on the benign end of the spectrum,
the phyllodes tumour looks like a fibroadenoma. And on the malignant
(10:15):
end of the spectrum, these tumours look like a metaplastic breast cancer. And
this is really problematic for these patients because if they
get misdiagnosed, say for instance, as a metaplastic breast cancer, when they actually
have a malignant phyllodes tumour, then they'll receive a form of
chemotherapy that is going to be completely useless on a
phyllodes tumour. The phyllodes tumours have no treatments available to
(10:37):
them at the moment. So it's really important that they
get the diagnosis correct so that they don't have these
incorrect treatments given to them.
Dr Viviane Richter (10:45):
Your research aims to prevent women from having to go
through this unnecessary treatment.
Dr Braydon Meyer (10:50):
Exactly. Yeah. So we're just trying to stop them from
being treated inappropriately because, on the benign end of the spectrum,
these patients, they'll have a slightly different surgery that occurs.
So we need to make sure that they're getting the
appropriate surgeries, making sure they are not being overtreated. They're
not receiving therapy that they shouldn't be receiving. The great thing
about this work, there's a functional aspect to this, so we can learn what
(11:12):
genes are potentially being disrupted in these specific tumours. We
know very, very little about them, and the more that
we do learn about them, the more we open up
that Pandora's box, where we can potentially lead to the
first therapeutic intervention for these tumours as well.
Dr Viviane Richter (11:28):
So, Braydon, how would this work in the clinic?
Dr Braydon Meyer (11:31):
So essentially, we're hoping we can take this signature that
we've found and develop it into a test utilising technology
that's already available in pathology labs all across the country.
We're kind of hoping that this test would simply be
able to be done in less than a day, and
it would just give this binary answer. It's going to
be either a phyllodes tumour or a fibroadenoma, or a metaplastic breast cancer in this particular case.
Dr Viviane Richter (11:55):
Can you talk to us about how you collaborate on
your research?
Dr Braydon Meyer (11:59):
So Amanda and I, we joke that between the two
of us, we're one functional scientist. Uh, when you drill down a little bit deeper though, we're definitely, like, very individual.
But in a really good way. I am what they
call the dry lab. I'm just doing a lot of
data analysis. I'm the one generating these models, these AI models that we're
talking about. Whereas, Amanda is more in the wet lab
(12:22):
and it's actually the interaction that makes it so much
more powerful because you don't want to be a jack
of all trades and a master of none. It's much,
much better to be an expert in one field and
then pass on your results to somebody else who is
an expert in their field.
Dr Amanda Khoury (12:37):
Yeah, Braydon sits in the office behind me, and he
bugs me throughout the day.
Dr Braydon Meyer (12:41):
All the time.
Dr Amanda Khoury (12:42):
All the time. Yeah. So my job is to come
up with the experimental designs, design the really complicated, you know,
what parts of the genome are we going to target? Can I
sequence some DNA to touch this bit and not that bit?
It's very intensely complex. And then I do those, what
I like to say, perfect experiments. And then I'll pass
it on to somebody like Braydon, who picks it up
(13:02):
and gets to interpret the results.
Dr Braydon Meyer (13:04):
And it's an iterative process as well. Amanda can pass
that data to me. I'll do my analysis, which might
then feed back to Amanda that instead of, okay, we
started with this whole genome analysis or something like that.
And now we're down to, okay, we have, like, three genes that
we need to look at. I give them back to Amanda.
She looks at those three genes experimentally. She tries to
(13:24):
knock them out, tries to figure out what it is
that their function and how they work. And then again,
she can pass that back to me. I'll do the
data analysis on her experiments.
Dr Viviane Richter (13:33):
How did you guys end up at Garvan?
Dr Braydon Meyer (13:35):
So my interest in science from the very beginning has
been epigenetics. I did it in my honours. I did
it in my PhD back in Melbourne at the Murdoch
Children's Research Institute. And then I had an opportunity, essentially, to
to come to Sydney and, yeah, work at the Garvan. This
was very specifically, so I could work in the lab
(13:55):
of Professor Susan Clark and Associate Professor Clare Stirzaker. Just because
those two are at the absolute forefront of epigenetic research
in Australia.
Dr Amanda Khoury (14:05):
Braydon's being a little bit humble here. Um, he actually
got offered a job in Singapore before he was offered
this job, but he turned it down because of how
much he wanted to work in our lab.
Dr Viviane Richter (14:14):
Amanda, where did your love for epigenetics begin?
Dr Amanda Khoury (14:18):
So I'm a bit of a latecomer to science. I,
in school, I studied the humanities, history, language, went to uni,
was doing a liberal arts degree, decided to try a
neuroscience subject. I thought that sounds kind of interesting. And
then that was where the 'wow' moment happened. I thought, oh,
my goodness. I've never thought about how powerful cells are.
(14:40):
Cells are what make you happy. Cells are what make
you healthy. Cells are what make you think. And I
thought these little powerhouses exist, and I've never thought about
them before. And I just sort of got obsessed with this
concept and then, literally, just did a full pivot into
a medical science degree. I really wanted to work at
Garvan because Garvan is one of the best research institutes
in Australia. And so for me, all I wanted to
(15:03):
do was get my foot in the door. I got
my foot in the door. I was working in this lab,
which was sort of doing routine drug testing for cancer.
It was interesting, but it wasn't stimulating enough. I wanted
to do more thinking about those mechanisms and what was happening.
And then I discovered the Epigenetics Lab, and I thought,
"that's it". And so then I saw Professor Sue Clark
and I said, "Can I please join your lab?" And
(15:24):
that's where the PhD began.
Dr Viviane Richter (15:26):
It's no secret that science can be a tough slog. Amanda,
what motivates you day to day? What makes you keep going?
Dr Amanda Khoury (15:34):
So I often think about the fact that you know,
I wake up, I get out of bed, I go
to work, you know, do my job, come back home.
It all seems really simple, but during the day there
are millions and millions of processes happening automatically in my
body that are allowing me to do that, and I
just find that really wondrous. And the thing is, these
processes are happening. But we're still finding out why they're happening,
(15:55):
what the mechanisms are. We still don't really know how
we tick. So what gets me out of bed is
really trying to answer those fundamental questions about why I exist,
how I exist.
Dr Viviane Richter (16:06):
Braydon, what about you?
Dr Braydon Meyer (16:07):
So, for me personally, I'm fascinated by rare cancers, and
there's quite an inequity in research across the globe for
certain cancers and certain diseases. There is absolutely no work
being done, for instance, on the phyllodes tumour or a
bunch of other rare cancers as well. But just because
these cancers are rare and they affect a very small
(16:28):
proportion of the population, doesn't mean that they shouldn't have
treatment options available for them.
Dr Viviane Richter (16:35):
Amanda, what is a big thing that you would like
to achieve in your career?
Dr Amanda Khoury (16:38):
So I've spoken about this idea that we've sequenced the
human genome, and then we thought that would hold all
the answers. And then we discovered the fact that the
DNA is actually not even linear. It's got a 3D structure.
But that finding only happened maybe 12 years ago. And
so for me, I'm absolutely fascinated by the three dimensional structure.
So I really want my role to be about solidifying
(16:59):
that structure and understanding how it's broken down, not just
in cancer, but in every disease. You know, if we find these
common mechanisms, that we could potentially find a common therapeutic target.
The really important thing about the research that Braydon and
I are doing is that the discoveries we're making, the
methods we're creating, they are not just applicable to single cancer types.
They are applicable to a broad range of cancers. So
(17:22):
what this means is we are doing research smarter. We
are combining AI, we are combining collaborative efforts, in order to really
find the fastest, smartest way to get more therapies to
more patients.
Dr Viviane Richter (17:35):
Amanda and Braydon, before we let you get back to the lab,
it's time for the Fast Five. Amanda, what was your
first job?
Dr Amanda Khoury (17:42):
I had the worst first job ever. I was the
woman in the perfume department handing out the samples.
Dr Braydon Meyer (17:49):
Look at you now.
Dr Viviane Richter (17:50):
Still very fashionable.
Dr Amanda Khoury (17:52):
Oh, thank you.
Dr Viviane Richter (17:54):
Do you have a favourite movie?
Dr Braydon Meyer (17:55):
I'd say my favourite movie would be Across The Universe.
It's a Beatles musical.
Dr Viviane Richter (18:02):
What is the current book you're reading?
Dr Amanda Khoury (18:03):
I'm currently reading two books. One is called Caledonian Road,
which I'm absolutely loving. It is about class inequality and
art and philosophy, and I'm also reading a Gentleman in Moscow,
which was recently ranked by The New York Times as
the third best book ever.
Dr Viviane Richter (18:19):
Something that's on your bucket list?
Dr Braydon Meyer (18:21):
My bucket list is I want to, like, step on
every continent. The major one being Antarctica. I'd love to go
and just hang out with penguins for a bit or
see the northern lights. That's a big one for me too.
Dr Viviane Richter (18:34):
Who inspires you the most?
Dr Amanda Khoury (18:35):
My parents. So they came to Australia to escape the
civil war in Lebanon, and my dad was the eldest
of nine kids. He couldn't go to school because he
had to just start doing labour jobs to support his
whole family. My parents raised a family of six kids
on one income, and we're now all university educated. So it's
a real step up. So they're my inspiration.
Dr Viviane Richter (18:57):
Doctor Amanda Khoury. Doctor Braydon Meyer. Thank you so much
for joining us on Medical Minds.
Dr Amanda Khoury (19:03):
It's been so good being here.
Dr Braydon Meyer (19:04):
It's been great. Thanks for having us.
Dr Viviane Richter (19:06):
If you'd like to know more about this research or
donate to the work we do at Garvan, head to garvan.org.au.
And if you've enjoyed this podcast, please leave a review
and share with other podcast lovers. I'm Dr Viviane Richter.
Thanks for listening.
This podcast was recorded on the traditional Country of the
(19:27):
Gadigal people of the Eora Nation. We recognise their continuing
connection to land, waters and community. We pay our respect
to Aboriginal and Torres Strait Islander cultures and Elders past,
present and emerging.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.