Subscribe
Share
Share
Embed
Proteomics in Proximity discusses the intersection of proteomics with genomics for drug target discovery, the application of proteomics to reveal disease biomarkers, and current trends in using proteomics to unlock biological mechanisms. Co-hosted by Olink's Dale Yuzuki, Cindy Lawley and Sarantis Chlamydas.
Welcome to the
Proteomics in Proximity podcast, where
your co-hosts, Dale Yuzuki, Cindy
Lawley and Sarantis Chlamydus from Olink
Proteomics talk about the intersection of
Proteomics with genomics for drug target
discovery, the application of proteomics to
reveal disease biomarkers, and current
trends in using proteomics to
unlock biological mechanisms. Here we have
your hosts, Dale, Cindy, and
Sarantis.
Hey, everyone.
Welcome to Proteomics in
Proximity. Today we have a guest,
Chris Whelan, who's joining us from
Janssen Pharmaceuticals.
Chris is the one who has
helped spearhead bringing proteomics
into the UK Biobank. So we're super
excited to talk to him about his history,
his background, and what
the vision of bringing
proteomics together with the
genetics that UK Biobank is so famous
for, the genetics and clinical data that
we're all very excited about on the UK
Biobank Research Analysis
platform. And this week is a pretty
auspicious week because we've just heard
that the first tranche of data from the UK
Biobank Pharma Proteomics
Project have become available
through the Research Analysis platform. So
we're excited to talk to Chris about that as
well. So, welcome, Chris. Hey,
Cindy, Dale, Sarantis. How are you all
doing? Doing great. It's great to
have you with us today.
Welcome,
Chris.
So Chris, can you tell us a little bit about
your background in terms of going into
science? You don't have to start sort of in
your elementary school days, but certainly
sort of your path
to industry because I think that's always an
interesting place to start.
Absolutely, yeah. Happy to.
So I did all of my training
up to getting my PhD in
Dublin, Ireland.
I've been told recently that I'm losing my
accent, so I'm going to try to make an
effort to sound more today.
But yeah, I
started off in psychology for my
undergrad and then realized I wanted to get
more into the cellular
sort of, sort of hard
science behind brain
illnesses. So did my Masters
and my Ph.D. in neuroscience.
One of my advisors was a
geneticist. So I started to dip my
toes in statistical
genetics. And that sort of
led me towards my postdoc in
Los Angeles with the ENIGMA
Consortium at USC. So there they were
combining neuroimaging with
GWAS, basically running,
genome-wide association
studies on very large collections of
MRI scans. So I did that for
two years. I
felt that I always thought that I would be
on the academic track. I remember
in my PhD class, they actually
wanted to do a straw poll of who wants to
go to industry and who wants to be a
lecturer or a professor. And I was firmly
in the professor camp. But
I think
two years in
academia in the States, it's tough.
It's tough. And I actually had a good
postdoc. My P.I. was awesome. Really
lovely, man. Really supportive. But it just
gave me some insights into it. It's
a tough place to be.
Beyond that, I think
I wanted
to be closer to the patients. That might
sound like a little bit of a cliche, but I
wanted to be really working on whatever I'm
doing. I can see this affecting a
patient in six or seven years
time. So, I was going to
move home to Dublin and then I
got the call out of nowhere from
Pfizer. And they were looking for
somebody who had a dual
neuroscience and genetics background.
So it just seemed too perfect to -
Ahead of your time,
Chris. So when they are
pulling those GWAS traits out of the imaging
data,
how is that being tracked?
What were the
connections you were looking for with the
genetic data? How were
they identified across different
MRIs that allowed it
to be
compared between
cases and controls? This
imaging area has
evolved so much
since I was in graduate school, so
I'm really curious how you did that.
So it's interesting, I think ENIGMA was
almost like a proto-UK
Biobank. I think UK Biobank
was in the middle of recruitment when
ENIGMA started up. But really it was a great
idea from Paul Thompson
where there
were a lot of different sites doing
MRI scans in maybe 50 cases or
50 controls, and
reporting differences in brain
structure and function that
sometimes were replicated and sometimes
weren't. So the broad sort of
oversimplified idea of ENIGMA
was, well, we can't bring everything
together, we can't ask everybody to just
throw their data in a sensor
repository. Ethics
and paperwork nightmare.
But we could agree
upon a standardized set of
protocols
to process the imaging
data and ask everybody to process it
in exactly the same way. And
then they all send us their
results because that's clean, it's
anonymized, and we'll meta analyze
all of our results together. So that's
where the name comes from
enhancing neuroimaging
genetics via meta analysis. Uh,
nice. Yeah, good memory.
I have to
type it out a lot during my
post doc. That's good news, that means a lot of publications.
How did the
UK Biobank come into your
life? How did you make a connection with
UK Biobank? And I think
you have also seen all the progress,
right?
Definitely, yeah. It's interesting,
I think that UK Biobank came into a lot
of industry
scientists lives around the same
time. While I was at
Pfizer, we were using
large-scale
genetic databases to
make inferences around
this gene is associated with this disease.
Maybe it would make a good therapeutic
target. But UK
Biobank came along, I guess around
2016,
2017. It really started to come onto our
radar when the exome
sequencing of UKB was announced.
That was one of the first sort of
major industry-
academic collaborations where
UKB worked together with Pharma
to generate the
biggest exome sequencing study ever
conducted.
That came on our radar as around
2016, 2017, I think
for me personally,
I moved to
Biogen in 2018. It
was around the time that Pfizer
pulled out of neuroscience and
Biogen were all in on
neuroscience. So it just seemed like the
perfect place for me to work.
But the first thing I was tasked with
when I joined Sally John's
organization was make
UK Biobank useful for
neuroscience, for Multiple
Sclerosis and Alzheimer's disease and
depression and Parkinson's,
et cetera. And it
was a sort of a tall order. I mean, UK
Biobank is breathtaking in terms of its
depth, and
it's just a beautiful, beautiful
study. But it's not a disease-
specific study. A lot of
these diseases like Alzheimer's,
they only come along when you hit your
60s. So
there's not a whole lot of people in
there, or there weren't, at least back when
I started working with it with
Alzheimer's. So there was not that many
questions that we could address using UKB.
So the lowest hanging
fruit for me, coming from my background
with the ENIGMA Consortium, was to look at
the MRI scans in UKB. Unlike
ENIGMA, which was retrospective,
metaanalysis, UKB
are actually collecting
scans across three different
sites in the UK, all using the same type
of scanner, the same head coil. It's
all standardized.
So that was the first thing I did. I did
GWAS and a couple of new
imaging measures from the brain
scans. Things like
local folding.
But
I felt like we could do more
to help neuroscience. And I started
to play around with the idea that maybe we
could look into doing
neurofilament light polypeptide
or neurofilament light chain in UKB.
So this is like a neuronal
cytoskeletal protein.
And when there's
some injury, when you get neuronal
injury, it gets
secreted into fluids.
So CSF [cerebrospinal fluid] or
blood. And it
was proposed, it was really
gaining momentum as a
potential biomarker for
MS [Multiple Sclerosis] and Alzheimer's and other brain
illnesses. So it just
seemed like a really exciting idea. What
if we could measure neurofilament across
UK Biobank, across these half a million
people, and we could get a database
of how much brain injury do you
have based on a blood sample?
But quickly realized that was going to
be very expensive and a little
bit niche as well.
There's not that many
pharmas that are invested in
neuroscience these days. And we felt
that if we were going to do it, we
would need it to be a multi pharma
consortium effort given its
expense. So thought about
it more and more and I
had already worked with Olink
on
smaller scale studies. What year was this
about? This was
2018, I
think. 2018, 2019.
But
I had been working with Olink
on some smaller scale studies. I'd done some
work in a Swedish neurology cohort
looking into proteomic changes in
Alzheimer's disease, and started to
talk to Evan Mills at Olink
about, "Hey, are you going to get
neurofilament on Olink any day? I'd
love to look at neurofilament in UK
Biobank." And we started to
toss around the idea that
maybe, instead of just doing
neurofilament in UKB, we could do
Olink because it captures
neurofilament and it captures many other
proteins at the same time. So we
could not just make this about
enhancing the value for neuroscience in UK
Biobank, but just in general,
enhancing the value for drug discovery and
potentially opening this up to a
wider consortium of
pharmas. But, yeah, that's a mouthful.
Basically, I can't remember the question
he asked. I asked you how
you got started with the UK Biobank.
And it's great because you zoomed right
into sort of getting 13
pharmas together. That was
no mean feat. What was it like? I mean, here
it is. You're going from one protein,
realizing that NFL is not going to be
of general interest, and then
some exposure to Olink. There
must have been a lot of different
conversations.
Yes. I don't know where to
start. If someone walks
up to you today and they say, how did you do
it? How did you make that happen? What
do you say to them? Because you shared with
me once that was a question you
get asked a lot.
Yeah,
it was a convergence
of factors, I guess, so to
speak. I think it was a
mixture of it was good timing because
I had been involved on the
Exome Sequencing Consortium, which was
eight different pharmaceutical companies
funding that, and that was wrapping
up. And we had a
conversation amongst the eight of us of,
what would we like to do next? Do we want to
do something next? And we basically took a
straw poll of other
multiomics techniques and proteomics
really rose to the top. So
I saw that as an opportunity. I was a huge
fan of proteomics to make my pitch
to that group of
companies. And it seemed
to go down well, but the
timing just happened to be right
because the field of proteomics was
maturing to the point where these multiplex
technologies could capture quite a
sizable proportion of the
canonical human plasma proteome.
And it just happened to be a time
where the
pharmas had budgets set aside to do
something innovative like this.
But yeah, and had a good network of
people helping me out. Melissa Miller
from Pfizer was a huge proponent of
this alongside me, and Lyndon
Mitnaul from Regeneron
as well. So lots of different people,
just basically all coming together and
agreeing that this was a good idea. I have
to just tell you that I was
talking to someone
on a different interview, and I said
Melissa McCarthy, because Mark
McCarthy and Melissa
Miller were both involved in
this. I just made that connection just
now, as you said that. That's funny.
Now, timing wise, you mentioned that you
started talking 2018
2019, if memory serves
correctly. I think there was a press release
at the end of 2020
announcing the UK Biobank's
involvement. So that must have been a very
busy year and a half.
Yeah, I've always had these bags under my
eyes, but they got bigger
in 2020.
The
first proper conversation that we
had about this was in
Pfizer's New York
campus in, I think,
May. Sorry,
February or March, I should say, of
2020. And I gave the pitch there.
And yeah,
then everything shut down. The whole world
shut down. So, the rest of the pitch was
virtual. So originally we got six
of those eight companies signed
up, and then getting the other seven on
board was all
meeting people for the first time from
different pharmas that it was all
through Zoom or
Microsoft Teams or what have you. Do
you think that Zoom
was an impediment? Or do you think
it actually because some things,
oddly COVID,
and this push to
Zoom and teleconferencing
kind of ushered in
telehealth that probably brought us a
decade forward in using
telehealth solutions. I'm just
curious about your perspective on whether you
think that helped or hurt or was neutral.
I have a silly perspective
on this. I
like it. I actually thought it was helpful
for two very silly reasons. I think the
first is that I can be awkward
in person, and I'm not very good at small
talk. So Zoom is very
you get online and then you get straight
into it. I've seen you in action
and you do get straight into
it. You get things
done, and then
I'm short. I'm like, five [foot] seven [inches],
so nobody can see that on Zoom.
Those are two very valid
reasons. Cut out the small talk.
And
I just took us down a rabbit hole, but
I love it. You
mentioned about multiomics.
How will you see the value of using
multiomics in big cohorts like the
UK Biobank? And what is the position of
proteomics? How will you see Protonics
position in this multiomics approach?
Yeah, that's a good question. And
sorry if this sounds a little
rehearsed. It isn't. But I've given so many
talks on this at this point that I'll
probably say the same thing that I often do,
which is that we've been using
UK Biobank and FinnGen and
these big population biobanks
to make links between
gene variants and diseases, and then
turn those links into
new drugs. So gene "X" is a really strong
association with disease
"Y". Let's turn it into a new drug.
Let's make a small molecule or an antibody
that hits the protein
that's encoded by that gene. Now
that hits the protein. We're not
measuring the protein, and that's the issue.
We're doing GWAS,
we're finding lots of new genes,
and a lot of them are intriguing, but a
lot of them are very difficult to drug. And
a lot of the time,
the gene association that we've
identified, it's messy.
I mean, it takes a lot of downstream
work to pinpoint exactly what gene it
is. And oftentimes,
it's either not completely clear
or it's very pleiotropic, where it could be
affecting a lot of different proteins
or pathways. So,
really, I always thought proteins as the
missing piece between
genes and diseases in that
genetics guided drug discovery process.
The proteins, we could argue
about it about how much they represent drug
targets now that we have gene therapies and
siRNAs, et cetera, that don't necessarily
target proteins, but
still, especially for bigger pharma,
the vast majority of the drugs we're
making are targeting proteins. So let's put
our drug targets part and parcel
of that genetic drug discovery
process, and then we have the potential to
maybe reveal something mechanistic about
how that drug is acting as well.
Exactly.
Yeah. And from the
pharmaceutical
drug discovery angle,
they intuitively sort of picked this
up, meaning they accepted that
premise that we go from
genetic guided drug discovery to
gene, to protein, to
disease.
I hope that they liked
it. They seem to like it because they
invested in the PPP [Pharma Proteomics] project. But,
yeah, I think that
it wasn't a difficult
argument to make, because I think people
have seen there've been a couple of papers
from AstraZeneca and Abbvie and
others, and they've looked retrospectively
at their drug development pipelines.
And they've basically assessed,
okay, which drugs made it
to patients and which drugs failed, and then
which drugs had support from GWAS
or ClinVar association
and which ones didn't. And there have been a
couple of independent studies that have
shown that if your drug target has
supporting evidence from genetics, then it's
at least twice as likely to actually
succeed. But there's a
lot of unanswered questions there
that seems to be pretty good evidence. Yeah,
okay, let's use genetics for drug discovery,
but there's a lot of murky stuff in
the middle that we still need to figure out.
So I think that's where the multiomics can
help. And as far
as the Pharma Proteomics Project
being, frankly,
you can say
it's a pilot, right? Because you're looking
at one-tength the size of the UK
Biobank. You can also make the
argument that, well, something like this has
not been done at this scale before in terms
of looking at 1500 proteins.
Were you
pointing to other work that had looked at
circulating proteins in genetics
as far as mendelian randomization,
that kind of thing? Yeah, absolutely.
There's been a couple of big studies.
Claudia Langenberg is one of the pioneers in
this field. She's awesome.
Well, I didn't prepare for this. I'm worried
I'm going to leave people out. But there's
Claudia, of course. There's Kári
Stefánsson in Iceland with
deCODE [Genetics]. Yeah, several different ...
There's the SCALLOP consortium that we're
doing this at a, I won't say smaller scale
because they'd amassed quite a large
collection of Olink data, but
just based on the old panels. So kind of
90 proteins at a time. So there had been a
lot, a lot of precedence. This definitely
wasn't the first time anybody was doing
this. It just happened to be the biggest
so far. So their
appetite was whetted. In
terms of these smaller studies,
they knew that this approach could work
and therefore that was really a low risk
decision. Do I understand that
correctly? To a certain degree.
I think that there were two ways you could
have pitched this. You could have pitched it
to
geneticists or you could have pitched it to
biomarker experts or
proteomics experts. And
I felt that the pitch was easier to the
geneticist because genetics for at least the
last 15 years, if not
longer, is used to doing things at
very large scale.
You need to do things in tens and now
hundreds of thousands. And some of the GWASes
are now even in over a million now in
order to pick up the biology, in order to
pick up the gene variants that are
influencing your disease. So they're used
to doing things at really large scale. I
think that they don't need to be convinced
of that. I think the biomarker folks are
more about let's do it with precision. I
think that they still needed some convincing
that we could do this at massive, massive
scale. But do you think the NGS [next-generation sequencing] approach
help you to make your
pitch to the geneticists because it's an
NGS approach and maybe they are more
familiar with this approach? How was your
feeling? Yeah, I
think so. I think a lot of folks had
used Olink before,
I think using the prior sort of
method, the PCR-based method.
I think that we'd seen some good
quality based on those data and felt that
the jump to NGS would allow us to scale
up like this.
What is the next step from the UK
Biobank? What's your
ambition actually first, and what's the next
step of a UK Biobank?
Yeah,
obviously it would be great to do all
half a million. And I think that we're
talking about that. We're having early
conversations about whether that will be
feasible financially more than
sort of technically. I think that it's
starting to become technically possible, but
we have to talk about how much it would
cost. I think in the shorter term,
we're hoping, and I hope I don't
jinx it by announcing it here, but we
have received approval
to do a smaller follow-up
study in 2500
samples in the UK Biobank. And
these 2500 samples have
already been profiled using the Olink
Explore assay. But we're going to
do three mass spec-based
platforms from Seer and
Biognosys and Eliptica, as well as
SomaLogic and then we'll just have a
very comprehensive
characterization of the plasma
proteome in these 2500 people. And some
of these people would have had COVID before
they entered the study and some didn't. So
it's sort of like, let's try to
capture as much of the plasma proteome
pre- and post-COVID as we can.
That'll be so
interesting, I think, especially to see
how the complementarity of these
technologies
wins out in a big cohort like
this. What are you able to reveal
if Seer has
this vision to be able
to sequence the proteome, try
and look at things that aren't
targeted, whereas some of the others,
are - we go
after targeted proteins. And
then I think these mass spec technologies
are well established as gold
standards and have advanced
very far in the last few years in throughput.
Because you're looking at different
things, right? In terms of what kind of
overlap there is with the canonical
protein or versus sort
of splice isoforms and all the
variety of proteoforms. I mean, oh my gosh,
there's what, 400 different
types of post translational
modifications. I mean you can
just start multiplying numbers together.
When people
ask you, because they've asked me this,
Chris, how many proteins do you think
are there, including
proteoforms, what is your thought about
that? You can't
have a wrong answer because we
don't know.
It gets kind of mind
boggling to think about, because obviously,
without the proteoforms, you would expect
there to be 20,000 just based on the human
genome. But then it depends on how many
you can capture in blood. In terms
of how they're expressed in different
tissues, proteoforms, I don't know.
Whatever I say will probably make me sound
dumb. Especially in five or ten years when
they work it out, like 100,000,
maybe. Yeah, that's what I've said.
I think I read somewhere someone made a good
argument around that. Maybe it was Karsten [Suhre],
maybe it was Jochen [Schwenk]. I don't know. Someone
said that. It reminds me of the speculation
of how many genes were in the
Genome Project.
The numbers were all over the place.
I mean, who would have guessed it would be a
little bit less than 20,000? I mean, not
that many, right? A lot of people really
thought it was a much, much larger number.
100%. Yeah, exactly.
And then, as far as I understand
that impending - or I'm sorry, already
we've got released the data in terms
of the Olink first
1500 [participants in the UK Biobank] against the
50,000?
Yes. I think that they are on
the Research Access portal.
Now, don't quote me on that. I do not represent
UK Biobank, but I think that they are.
Naomi told me Monday, no told me
Friday that she said it was up there.
So by the time this podcast comes out,
I think you're pretty safe.
There's several weeks-long lag time here,
so we're looking at May
2023,
the first sort of set
of roughly how many
samples? It's probably about
54, maybe 52 after
QC. A thousand
samples. 52,000
samples times some
1469 or
so, give or take,
proteins analyzed by Explore
1536. I mean,
that's quite a data set
for people to dig into.
I think - go ahead.
No. Go ahead, Dale. Sorry. I was thinking
about all the posters at
ASHG [American Society of Human Genetics conference], right, in October that we were
talking about on the podcast, as far as how
many there was, what, 19 or
so abstracts of different types
of work. Yeah. This is not bragging. I
have to keep track of this so I can convince
people at Janssen and other companies that
this is a return on investment. But yeah,
you should brag. I think it was 19
abstracts and six talks at
ASHG. But what I'm really excited about the
public release is
that's obviously a lot of
output, especially for a data
set that's so new, but
I don't even feel like that's
not scratching the surface even. I think
there's going to be so much more that
academics can do. There's a lot of creative
things that you could do with this data set
that might not have immediate translational
impact for drug discovery, but academic
scientists are going to take this and
probably do something really revolutionary
with it. I can't wait till next
year's ASHG once these
publications start getting into the
literature, right? It's going to be all over
the place again. Is that because
there's such a wide variety of
different phenotypes that they can
associate protein level and genetics to?
Yes,
well, yes, to a certain degree. I think
we've looked at that. I think probably a lot
of the companies have looked at that. We've
done kind of an all by all. Take all of the
ICD [International Classification of Diseases] codes or the feed codes, and
then run a regression against
all the proteins. And that
basically gives you a crude biomarker
study. And we've been using those results
in house. But
I'm mainly thinking just about
how there's so much creativity
out there in the academic community. There's
questions that you could address with these
data that we probably haven't even thought
of yet, because this was
UK-PPP was like, one project
out of several on our plates and
pharma. So I think that
fingers crossed, the academic
community will have a lot of fun
ideas. Well, we
had pushed out
an Explore 1536 data
set when we first launched that Explore
platform, and it was on COVID. And
there have been publications spurred from
that by just comparing those
COVID data in that
cohort over to
whatever work the researcher was
already doing to look at those different
signatures. So
seeing publicly available data
spur novel comparisons
and novel publications. I
just think that's what
it's all about, right? Getting these
creative minds on it, crowdsourcing these
ideas and how people
debate ways to do things on Twitter,
I just absolutely love.
It's fantastic. On
UKB, I think it kicked off in 2006,
so it's not a new
study, but it always feels new. They're
always adding cool, innovative
new technologies to this data
set. So it'll go on for a long time to
come, for sure. And then as far
as you being how,
do I say that, that organizer. You were
there at the beginning, you must have
lots and lots of invitations to give these
kinds of talks. As far as UK
Biobank and the PPP in
particular.
Yeah, I do. Don't ask
why did he accept
ours?
No,
I do. Yeah, it's exciting, and I want to
make sure that I spread it around. I guess I
am the P.I. for the study
overall, but there's no way this ever
would have happened without a lot of other,
more talented and more
intelligent people than me involved.
So, I get invited a lot, but
when I can to try to forward along to
other folks who help build this
as well: Melissa [Miller], Ben
Sun,
Joseph Stakowski,
and by the way, Brad Gibson
from Amgen
was a huge proponent because
the consortium was
90 something percent
geneticists. Brad is the
proteomics expert. Brad has the real
background, the hardcore background
mass spec. So he helped put
guardrails on this and make sure that we
were doing things properly. Make sure
he got the ball over
the finish line, too, right? In terms
of that extra weight of
somebody who is not coming from the genetics
field, but within the pharma proteomics
sort of context.
I think so.
I have
probably
built a little bit of a reputation in this
study, but I didn't really have any
before I started it. And
I think when I was pitching this idea,
there was probably a lot of skepticism,
like who's this little twerp? And he has
his genetics background. So Brad
being on board and putting his
weight behind it. Mark McCarthy, as you
mentioned earlier, Cindy is involved as
well. And Carrie, there were a lot of
people that
are more prestigious than me,
put their weight behind it, and really
helped put it over the finish line. Well,
they got behind your vision. That's got to
feel good.
But how will you see - I have a question
now, generally more for the cohorts. How
will you see the use of cohorts
in pharma and the drug development
process? I mean, what is the value and
what do you think is having different,
also ethnicities and different,
let's say, from different
places the cohorts will help on that?
What is your vision? How's your idea
about that?
I think that they're
the engine for
sort of
epidemiological
health studies, basically
for any sort of common
complex disease, we need these
population cohorts to
gain a better understanding of their
molecular mechanisms, the causal
mechanisms, as well as potentially some of
the environmental influences on these
diseases. So, I think that we've
done a lot with UKB. There's a huge push
now, a very well deserved
push towards looking into
underrepresented population cohorts.
So lots of different ones that we
could potentially look into.
And also
disease enriched cohorts, cohorts
that might have a dementia
wing, for example. I know that Finngen
is building up its dementia substudy,
so lots of different
directions that could go in. Is there a
threshold, a minimum threshold,
maybe because of incidence of disease in
these cohorts or something? Like do you
not tend to look at anything less than
10,000 samples or anything
less than 50,000? Or do you look at
each cohort on its own merit, based upon
is there longitudinal
data? How are the data
collected? Can you share a few
criteria that you think are important for
selecting cohorts? That's
a good question, and actually, we've
started to think about this more
objectively. Can we put together a list of
criteria for biobank
curation? Because now that the UK
Biobank used to be the only game in town,
but it's still probably, in my opinion, the
best, but there are a lot of excellent
cohorts coming out as well.
The way that I think of it, and this is a
little bit coarse and
maybe crude, but
the larger your
sample size, the
less detailed your phenotyping
and your clinical information, and then the
smaller the sample size, the more
disease specific clinical phenotyping you
can get. So I would say you could
go all the way up to some of these medical
records databases from Optum or
IBM, and they've got hundreds of
millions of people. And you can do some cool
things with regard to
comorbidity mapping in those databases, but
you can't link to a specific
clinical scale for depression or
Alzheimer's disease and they're not going to
have neuroimaging or
proteomics, et cetera. And then on the
other end of the spectrum, you have some of
the cohorts that, like I mentioned at the
start, the Swedish Neurology cohort that I
was applying Olink to a few years ago
that's got CDR [Clinical Dementia Rating] summary
boxes and mini mental state examination
and all of these very disease
specific measurements that really
help us drive in
on specific
hypotheses that are relevant to disease and
sometimes almost use those studies like
natural history cohorts or like control
wings to clinical trials. And then you have
sort of the, I won't say the "Goldilocks"
biobank, but the goldilocks sort of
approach, but I can't think of a better
term. And that would be where the Biobanks
fit in, I think UKB
it doesn't capture everything, it doesn't
have many mental state examination or
CDR summary boxes. But it does
have fluid intelligence
tests, it has trail making. It has a lot of
different cognitive and
functional tasks,
paired with deep
genetic data. Now, proteomic data,
actigraphy imaging, et cetera,
et cetera.
So, I think that finding that sort
of goldilocks
approach where we can get the power of large
scale, but also get some of the
denser clinical phenotyping, is
usually how we try to go about it when we're
selecting our cohorts.
That's amazing. Wow. That was really
a rich answer. Thanks.
Well, Chris, it was great having you here
on the podcast this afternoon.
Thank you, really, so much
for being so generous with your time and
your thoughts today. We really look forward
to seeing some results.
And indeed, like you mentioned,
the creativity of scientists.
I'm very happy to be here.
Before I go, I do want to give a shout out
to Evan Mills again for helping get this
over the line and
to Klev Diamanti and
Philippa Pettingell, who did so
much technical and
just all sorts of technical support and
scientific support for the UKB
project. A disclaimer:
this was Chris's shout out to
three Olink employees in
recognition for all
their effort on this. But I'll also say that
I think Klev and Philippa have both said
how much they appreciated, how
much they learned from the
genetics perspective, from
so many of these thought leaders that are
the scientists in pharma who are
driving the experimental design and the
vision and gained
approval to use the UK Biobank
data. I think this idea of looking at
pharma as a funding body,
you shared that with me before,
Chris. These are heavy
hitting scientists that
have an
incredible track record of being able
to drive
such rich discoveries.
So it's such a
privilege to be around
you all and see this paper
coming out
from these data that
you've all been a part of. So I look forward
to that publication, too, in case
you can plug for it. I don't know if you
know any timing around the
UK-PPP paper.
First paper. I
resubmitted the revised version over the
weekend.
Someone else
was working over the weekend then.
The response to the reviewers
was 29 pages long. That can either be a
good thing or a bad thing
A lot of novel methods, I think.
Well, that's
exciting. That's exciting. You heard it
first here. Yes. Plenty to look forward
to. Thanks again very much,
Chris, it was great. Thank you.
Yeah, good to be here.
Thank you for listening to the
Proteomics in Proximity podcast
brought to you by Olink Proteomics.
To contact the hosts or for further
information, simply email
info@olink.com.