55:41
Methylation risk scores
01/15/2025
Season 2
Episode 18

Methylation Screening Array: A Revolutionary Tool in Epigenetics

DNA methylation, a cornerstone of epigenetic research, is vital for understanding gene regulation and its implications in health and disease. In this week’s episode of the Everything Epigenetics podcast, I speak with David Goldberg and Nicole Renke about the latest advancements in DNA methylation tools, including Illumina’s groundbreaking Methylation Screening Array (MSA).

We explore the history and evolution of methylation microarrays, why they remain the gold standard for Epigenome-Wide Association Studies (EWAS), and the design process behind the MSA array. David and Nicole provide insight into the practical applications of these tools for research in aging, neurodegenerative disorders, and environmental exposures.

In this podcast you’ll learn about:

– What methylation microarrays are and why they’re essential for EWAS
– The market need and vision for Illumina’s new MSA array
– Key features and benefits of the MSA array for researchers
– ⁠The role of cell-specific methylation in advancing epigenetic studies
– ⁠Insights into the technical and biological validation of the MSA array

About this Guest

David Goldberg & Nicole Renke

Illumina Website

Nicole’s LinkedIn

Nicole’s Twitter(X)

Welcome and Introduction

Welcome to the Everything Epigenetics podcast, where we discuss DNA regulation and the insights it can tell you about your health.

I’m Hannah Went and I’m the Founder of Everything Epigenetics. Today I am so happy to have on actually two guests with me

David Goldberg and Nicole Renke. Now this entire episode is going to be a little bit different.

We are really going to be focusing on DNA methylation tools, including a luminous,

very new tool called the Methylation Screening Array or MSA array.

This is so important and why I like to cover a lot of different aspects on my podcast, because without these tools, you don’t

get the insights that epigenetics is able to provide us, right? We don’t get the insights into disease diagnosis diagnostics.

We don’t get the insights into biological aging or other therapeutic implications that this information can offer us.

So we talk about Nicole and David and their background and how they got into this space. We’re even going to give a brief overview

about methylation in the methylation screening array. We’re going to answer questions like what are the applications of Illumina

as microarrays and why are they used today? What’s an EWAS and why are they popular for EWAS.

And then why do people even like to use methylation microarrays when we have methylation sequencing tools available?

We’ll also understand a little bit more about the MSA, the methylation screening array broadly, and the market which needs these types

of tools, including, again, researchers, diagnostic companies,

even all the way down to the clinician and patient themselves.

So we even talk about how this array was designed. David walks us through the content selection

process, explaining all the sources of content that were used. And even talking about the new updated content this, this tool has on it.

We then end by talking about what is the most exciting thing about the methylation screening array, and even the preprint

that David and Wanding played a vital role in. Finally, we end on what other methylation products are in development

and what Nicole can tell us about them. Now, a little bit of an introduction into my guests.

Nicole is a Global Product Manager of the methylation arrays at Illumina. She works with a lot of teams there to really help

History of Methylation Microarrays

bring new technologies to the DNA methylation microarray business. She joined the Infinium methylation product management

team in 2020 and has led a lot of these array launches. She received her Bachelor of Science and Master of Science degrees from Tulane

University and holds certifications in product management, marketing, and DNA methylation. Analyzes.

As for David, he is originally from upstate New York and received a B.S. in Biological Sciences.

And then he was really into researching neurogenesis in the adult mammalian brain.

He’s actually currently a PhD candidate at the University of Pennsylvania in the lab of Doctor Wanding Zhou, where he is

developing computational methods for the analysis of DNA methylation and studying it in neurodegenerative disorders.

David has worked with Dr Zhou to develop the content on the methylation screening array Illumina.

His most recent array, which we talk about today. And he’s currently using this platform now to study the dynamics of DNA methylation

and hydroxy methylation across diverse human tissue types.

I will definitely be having both of them back on separately on the show, just because they’re so interesting,

and I’m personally just interested in, you know, their, their day to day work and how they’re moving the epigenetics field forward.

Now for my guests, David Goldberg and Nicole Renke, I hope you enjoy.

Welcome to the Everything Epigenetics podcast, Nicole and David, I’m excited to have you here today.

Thanks for spending your time with me. Thanks for having us on the podcast, Hannah.

Yeah, it’s great to be here. Thank you, Hannah. Yeah. So I’ve recently been doing more podcasts with more than one person.

It’s super fun because we get a couple different viewpoints. You know, Nicole and David, they each have their unique skill set.

I just gave you a little bit of background about them, but I want to hear it from them too. So we’re going to kind of ping pong this.

So I’m going to give some questions to Nicole, then go to David and keep going back and forth.

Nicole, you are the Global Product Manager for Methylation Arrays at Illumina.

I’ve spoken about this topic several times on on the podcast before, but from you, what does this role entail?

What does it mean? What do you do on a day to day basis? Yeah. So, it’s a really exciting role,

and I’m really grateful to be in this position. Basically, what I do as a product manager is,

I do about 50% of my work involves kind of product strategy, and the other 50% is more on execution.

So bringing new product ideas we have to the market.

So as sort of the business lead of DNA methylation microarrays,

I work with the development teams to bring new products to market. Like I said, we also work with operations teams, marketing and sales.

So really every step of the product life cycle and, what’s really important in product management

is to have a really good pulse on user needs.

So, what I try to do is stay very close to the market, you know, not only meeting customers in person,

but also reading a lot of papers, staying up to date with the literature and DNA methylation,

and bringing those insights into our product development process. And again, I’m just

a product manager can’t really do anything without, you know, all these all these different roles that,

support, so we’re one we’re one piece of the pie, but I’m just so humbled really all the time to work with

just brilliant scientists on the product development side. And just really great people who are on sales

and marketing and actually, you know, get out into the field and find the right sorts of users for our products.

Yeah. That’s great. I have a new profound love for product teams. I never knew what that was, what that meant.

I’m like, what, what are—the what does a product like person product manager do? And we now have a full product team at TruDiagnostic,

and it is like made my life so much easier. So you’re like the glue that yeah holds everyone together.

So I love that we have you here. Speaking, from Illumina on kind of the creation of this, this new methylation

screening array, we’ll, we’ll get into and then we also have David here. He’s from University of Pennsylvania.

David, what has been your research focus of your PhD, and how did you come

into this world of epigenetics, specifically DNA methylation?

So I’ve been interested in DNA methylation for a long time now. I studied biology and I just always found it fascinating

that we have, you know, trillions of cells in our bodies, hundreds of cell types,

very different organs and organ systems, but they all have the same, DNA.

And it always just fascinated me how you could get such different cell types and, tissue functioning from the same genome.

And a lot of that is regulated by DNA methylation. So it always really interested me.

Some of the early studies, using mouse models where they showed that, you know, depending on the diet that you gave,

the agouti mouse, it would have different fur color, different, predispositions to metabolic disorders.

And this was all regulated by the methylation at certain genes, which was a direct readout of the environment in the diet.

So I always found it fascinating. Really going from that molecular level

on top of the genome to the actual phenotype, these days, during my PhD work,

I’ve been doing a lot of research in neurodegenerative disorders. So focusing on Alzheimer’s disease

and primary age related to apathy and trying to use DNA methylation,

to predict Alzheimer’s or the level of tau in the brain,

using just the marks on the genome, which has been, really interesting journey.

And then on the other side, I also work in some tool development. Obviously the development of the MSA array,

as well as software program to facilitate the functional enrichment

testing of CPG queries and understanding what,

certain cytokines are enriched in, throughout the genome. To better interpret results from EWAS studies

and other methylation based analyses. Yeah, definitely. Well, thank you know, for that that little bit of an introduction.

You know, I had Randy Jirtle on my podcast and we talked about the agouti mouse mice trial

and, you know, what he’s done with his work in the imprintome. So that was like, you know, a dream having him on.

I remember looking at that when I was a little girl studying biology and being like, this is so cool. So, that’s great.

And then personally, I’m interested in Alzheimer’s as well. I have an APOE4. So, you know, being able to detect even risk,

from the DNA methylation level and epigenetic level and then maybe being able to to reduce that risk with set therapies,

I think is a personal interest of mine too. So, so yeah, we’re going to hop into, you know, kind of this, this methylation screening array.

We’ve said this a couple times now, that’s really the, the main point of this, this episode today, you’ve heard me talk a little bit

Applications of EWAS in Research

about kind of these microarrays or these bead chips we call them. That’s what we’re physically running your sample on.

At least when you send one to TruDiagnostic when we measure your epigenome.

So, Nicole, this these next set of questions are going to be, you know, directed toward you.

Can you just give us a very simplified definition of what is a methylation microarray?

And then also maybe a little bit of the history behind them at Illumina.

Yeah, definitely. So in short, a microarray we can think of generally

as a laboratory tool that we use to measure DNA methylation and

to kind of explain it, I think I’ll go back to just the general principles of DNA sequencing.

So we have in DNA sequencing, we are basically trying to decode the genome

and read the sequence of DNA nucleotides, and we have technology to do this,

and it’s been around for several years. The way this works is,

in the simplest terms, the DNA nucleotides are fluorescently labeled and we use a scanning device,

whether it’s a sequencing instrument or a microarray scanner. The function of these devices is to actually pick up

on those fluorescent signals. Basically, they’re colors. And, what this— what the color signal tells us

is the DNA base at a given location in the genome. Now for methylation, what we are trying to do

is not to measure the DNA nucleotides or those DNA variants, but we are actually trying to distinguish between methylated cytosine

and unmethylated cytosine. So primarily, and just to give people some context,

when we talk about DNA methylation, the most common type of DNA methylation that’s been the most

well studied is DNA methylation, which is a chemical tag

that basically attaches to the cytosine base specifically.

So, what we do is, to measure methylation, we actually have to convert

the DNA sequence to a different format that really allows us to interrogate methylated cytosine and unmethylated cytosine.

But it’s still based on that found that foundational technology where we are using these fluorescently

labeled nucleotides to pick up on, the methylation signal at a given location in the genome.

And so, in short, that’s that’s how we measure DNA methylation is we we use this foundational technology,

but we do an additional step upfront to convert the genome, and that allows us to figure out the methylation,

signal at a given location. Yeah. That’s a really great definition then.

I know these have definitely evolved over time. Right? Any of the researchers listening?

You know, you go back and read research papers and there’s what I think it’s a 27K array.

Right, a 450 k array etc.. Do you just want to give a little bit of,

a history of maybe the growth in the expansion we’ve seen how really far we’ve come with this technology?

Yeah, it’s really interesting because the timeline is actually quite long and,

it’s been really like, a privilege to work with all these researchers

who have used our iterations, you know, the very first iterations of our methylation microarrays.

And when that started was actually an assay called GoldenGate, which launched in, I believe it was 2008.

And this technology could target around 1500 of these CPG sites.

So these cytosines—where methylation is, typically known to happen in the genome.

So 1500 CPG sites is where we started. And then in 2010,

we had the 27 K Infinium array. So that targeted 27,000 CPG sites.

And back then it I mean, you know, I was not a researcher back then, but it must have been something that was really,

profound to be able to measure that many more. And, another big advancement

was the 450 K array, which was launched in it was launched in 2013, and

that covers 450,000 CPG sites in the genome.

So now we’re up to half a million. And this was really the tool that,

a lot of the field of EWAS, which is Epigenome-Wide Association Studies was based on.

So there are thousands of data sets that you can find that are publicly available on GEO that,

have for 50 K array data. And what was the next

advancement was MethylationEPIC in 2015. And this pretty much doubled the content of the 450 K

array up to around 860 thousand CPG sites in in the genome,

and the content was improved. So we were learning about, you know,

what were the important areas to cover in the methyl. And this content really expanded the coverage into

not only the promoter regions, which were very well established at that time,

but also getting into new areas like enhancers and intergenic regions.

And then in the last few years, we’ve also expanded the product portfolio beyond humans.

And we launched a mouse methylation beadchip in 2010. So that is very useful for researchers who are studying model organisms.

And, that is around 285,000

CPG sites and can sort of be considered an epic for mice. And in late 2022, we came out with EPIC version 2,

which has some expanded content that covers, more enhancers, super enhancers, CTCs, binding sites.

And like you mentioned, Hannah, what we’re really excited about these days is our new methylation screening array.

And this I really had the pleasure of working with, David and Wanding, from the Children’s Hospital.

Philadelphia. And, this tool is really designed

to be the ultimate EWAS and, tool that, I believe is going to take us into the next generation

of epigenetics research and really lead us into applying epigenetics

in the population health context to very large screening cohorts.

So for MSA, we were really focused on building a tool

that was a lot more scalable, a lot more affordable, and had the right content selection because this is a targeted tool.

So our priority with the content was kind of twofold. We wanted to enable researchers to be able to use the same powerful,

CPG sites that have been discovered to be associated with a wide variety of traits and cell types

from the whole history of, of EWAS and all of the literature that’s been published today.

And we also wanted to enable new capabilities. So new biomarkers that have been

that so new methylation sites that could be very useful for research.

And this is based on, you know, new research cell type outlooks and new sort of,

whole genome bi sulfite sequencing data sets and databases

that have been built up since epic was launched in 2015. So to kind of execute this vision, we we had to outsource

because this is something that, we knew,

we, we needed some bioinformatics expertise on. And, and this is where Wanding and David came in.

So Wanding and David has,

Wanding’s lab has been, very instrumental in previous the previous designs of Infinium beadchips like the mouse methylation array.

And, so we reached out to them and, they were really the content leads

Unveiling the MSA Array: Vision and Design

to help bring this vision we had to life. And we’re just so grateful to be able to work with them

and just they’re… really appreciate their, collaboration ability.

So not only did they kind of look through all the literature, but they also reached out to other users and we’re just very collaborative.

Yeah. Thanks for that history, Nicole. That’s awesome. And remember, you all listening like this is the tool Nicole

and David mentioned at the beginning to measure DNA methylation. So without this tool,

we can’t have all of this awesome science and research coming out of this. So, you know, there are a ton of clinical implications of DNA methylation.

And we know, you know, DNA methylation patterns are altered in all types of different disorders.

You know, David mentioned he’s studying more like, neurodegenerative, neurodevelopmental, maybe disorders.

There’s, you know, epi signatures as potential biomarkers. So being able to even report out other,

like clinical lab values or other parts of the multi through DNA methylation as well as other clinical implications and further research

and therapeutic implications. They’re obviously huge in the epigenetic aging clock

space, which I talk frequently about, even all the way to environmental exposures, methylation

risk scores that I spoke with my, Michael Thompson out of UCLA on. And then what Nicole just mentioned to write these studies

so that the methylation and, array, that the screening array, the MSA array we’re talking about today, I think can really help EWAS,

and we’ll talk about, how here in a second. But remember, those EWAS are just a really,

a type of research that measures population level differences and DNA methylation on a large, large genome-wide scale.

But it’s going to be methylation wide, right? Because we know DNA methylation is really the most stable epigenetic marker.

And, you know, technology for measuring other epigenetic markers hasn’t really been robust historically.

So I wanted to give, again, a wider viewpoint on why it’s so important to talk about this, this new tool in this new technology.

Now, David, this one’s going to be for you. You know, what is your perspective of of just

methylation microarrays in general as a user of them? Like, why do people still use these microarrays?

Why is Illumina coming out with all these new versions of them? When there is methylation sequencing even available?

So kind of a two parter, you know, how you use them as a researcher. But, you know, why do we even use them still over the sequencing too?

Yeah. So, there’s a lot of advantages to using the microarrays.

First and foremost, as Nicole mentioned, is the scalability.

For whenever you want to study a trait in the population, whatever type of disease or environmental exposure,

you typically want to profile a lot of samples, especially because, humans

have a lot of inherent variability, different ages and sex, things like that could impact your DNA methylation.

So if you really want to focus on one trait, you need to have a very large cohort,

and doing whole genome bisulfide sequencing at this time, it’s just not economically feasible.

So the arrays are great to really do these population scale studies.

And MSA is really going to be pushing that to even the next level in terms of the volume of samples that could be profiled.

Another major advantage is, the data is,

very accessible to work with. So, there’s a lot of tools because this has been

the main way of profiling methylation in populations. There are a lot of software packages out there that have already been designed,

and tested by the community for analyzing the array data.

And a lot of models have been developed as well. Like you, like you mentioned, many of these epigenetic clocks,

are very well characterized and in use now. So if you have a new data set

or you’re studying a different disease, you can apply these existing models, and see if there’s any sort of acceleration,

and what that model is reading. I think another important advantage is the

quantitative resolution. So when you do sequencing often times you’ll get parts of the genome.

But at any given CPG site, it’s very rare that you’ll have very high coverage.

Meaning you’ll have very few reads that actually have mapped to that individual CPG,

whereas on the array, some people have compared it to like 50 to 100 coverage.

Your measurements have very high precision and you can be very confident most of the time with the methylation readout that you have at those sites.

And for those reasons, you don’t have to deal with a lot of the data sparsity challenges that you have,

with whole genome bisulfide sequencing data, where you often have to merge the methylation over a genomic interval

and you start getting a lot more sort of course measurements.

And yeah, I mean, less computing power is required. Less memory and just really well-developed tools and pipelines.

So it’s a pleasure to work with and you can really profile a lot of samples.

So I’m a big fan and I think that… -Good, good. Now it sounds like it I mean it just sounds like there’s,

you know, many more advantages or it makes sense that we’re looking, or still using that these microarrays,

for specific use cases, I guess, depending on your area of research interest. So just to summarize to David, you said like scalability

costs and volume considerations, accessibility to the data. So it makes so much sense. There’s all these software packages.

There are all these tools that you can already use that have been validated within the community.

Quantitative resolution. You don’t have to worry about, you know, not having, a good enough or good amount of coverage.

The microarrays are really great at that. And then, yeah, less computing, right? Who wants to do more computing these days?

So it has, you know, less memory, you know, easier data to, to work with as well.

So thank you for that. Now, what we’ve been hinting at this whole time and what we really want to talk about,

this new methylation screening array, Nicole—Illumina has launched it this year.

Tell us more about it. I know why we’re so excited to be able to use it at TruDiagnostic.

So I’ll definitely add my perspective, but we want to hear it from you first. So yeah, walk us through the market.

Like, why did we need this? Why was this a product that, came to the table and became available?

Yeah, I’ll actually start by maybe adding on to, you know, the previous,

the previous points about microarrays and their advantages. So what we see today as an advantage of microarrays

and why they are still around in the context of studying methylation, is that they’re very reproducible and quantitative,

and they’re able to call methylation differences in a way that,

it or they’re, they’re able to call methylation differences in the context of EOS,

that can lend themselves well to a very powerful, very reproducible biomarker.

So there’s been, just thousands of papers that have been published,

on using DNA methylation microarrays to discover biomarkers that are associated with something like cancer or,

even environmental exposures like arsenic exposure, air pollution,

a whole array of diseases, as well as just natural phenomena like aging and,

so that’s and there is, of course, a role for, for DNA methylation sequencing.

And those tools are going to become better and better. Illumina is also investing in new technology

platforms for DNA methylation sequencing. And I think where this is really useful is to discover

sort of new areas of the methylation that haven’t been discovered before, kind of on a genome-wide level.

And understanding really at a deeper level, some of the biological pathways and how genetic

variants interact with these, whole regions of methylation

differences that we see between cases and controls. But, for methylation microarrays, they’ve really been a popular tool.

And over the past few years, we have studied how people are using these tools.

And what we found is that users really, we really have two different groups of users.

So one of the groups— group 1 is cancer

researchers and also researchers that are studying,

researchers who are studying rare genetic disorders. And what’s interesting about these types

of diseases and disorders is that the methylation differences

that you see between, on average, your cases

and your controls are actually very large. So oftentimes we see 20% or more

methylation differences between, see a

the methylation patterns of a, let’s say

a tumor sample. Say you’re studying lung cancer. And then you take a biopsy of the lung tissue and look at the DNA

methylation patterns across a whole variety. Across a number of people who have lung cancer.

And you compare those patterns with, you know, a biopsy of normal lung tissue,

you actually are able to detect methylation differences that with a high amount of power and a small number of,

individuals. So we actually only need small study sizes in certain

research areas. But the vast majority of our researchers who are using methylation

microarrays are using them to study common diseases. So like I mentioned, things like psychiatric disorders,

public health, aging, environmental epidemiology.

So, you know, exposures of, say, maternal smoking things like that.

So what the difference between these users is they actually need to have higher, a higher number of samples

to make those methylation biomarker discoveries. And that’s because the methylation differences we see in common diseases

or environmental exposures are actually quite small, often under 5%.

And in order to to find these robust biomarkers that you could say

use to develop a classifier, you really need a lot of samples.

And this is what we refer to these users as kind of EWAS research type.

And over time, we’ve heard that these types of researchers,

they’re actually looking to run studies of at least a thousand samples. And when they’re studying diseases or exposures

with even smaller predicted methylation differences, they actually need to scale that research to 10,000.

To 10,000 samples or even more. And at the same time,

we have a lot of national genomics initiatives that are popping up and have already done sequencing or genotyping of,

entire cohorts and are using these insights to improve population health and the field of epigenetics.

It’s really becoming very—

it’s gaining a lot of traction. People are now learning.

People now are interested in studying epigenetics and DNA methylation

is the most available type of biomarker to study within epigenetics. So we are seeing these big projects that have a need

for methylation profiling. So kind of a long explanation. But the methylation screening array was a tool that we needed that we knew

that we needed to be more affordable than the existing tools we have.

We knew that it needed to be more scalable. We needed a tool that would be able to profile, say,

a million, up to a million samples in in just a couple of years.

And, we needed a tool that had content that was going to be very valuable

for us researchers and for the study of diseases.

That would be considered under the umbrella of population health. Yeah, yeah. No, I know that’s comprehensive.

And you said, you know, it’s a longer answer, but I think it’s it’s really good to to touch on those points. So, you know, from, from the laboratory standpoint at TruDiagnostic,

I mean, this is more affordable than we can pass off that cost to, you know, our end consumer and, and start to make

these diagnostic based tools much cheaper to the end user. I think that’s really the goal is being able to, get these

Technical Validation and Biological Insights

into everyone’s hands, not just people who can actually afford it. So by you all doing that, it helps us.

It helps, you know, the end user, the public, as well. It’s super scalable as you mentioned.

So, you know, we used to only be able to run eight samples on one of these microarrays.

And now we can run 48, which is amazing. So it allows us to have a much more higher throughput in our laboratory.

And then you mentioned the content too. You know, we have defined content in terms of CPG locations

and particular, areas of the film that we want to look at. So, being able to have that defined content and maybe a more concise,

you know, microarray, it just makes our lives a lot easier as well. And that’s the next point I really want to go to is, is that content piece?

David, I know you played a massive role in the content design of the methylation screening array, the MSA.

So I’d love to hear a little bit more from you, how you and Wanding actually chose to that the content.

So feel free to walk us through the selection process. All of the sources that were used, I get this question

a lot like who who comes up with this stuff, right? Like who picks what we’re actually looking at.

And, you know, I have him here on the podcast, so you all can can hear from him directly.

Yeah. So, well, fortunately, a lot of people, really helped with the content selection

because we really leverage the thousands of studies that were done.

And, a lot of the prior probes that were from previous platforms. So basically the array was designed in two phases.

The first half was trying to identify previous probes from Infinium platforms.

That had a lot of associations were statistically robust, and associated with the diversity of traits.

So essentially what we did was we curated thousands of studies from EWAS databases,

a lot of manual curation. Talking to people in the fields that are doing population scale, EWAS studies.

And what we did was group them into approximately 20 different trait categories.

So we were able to use some, natural language processing to group them into different traits.

And these range from cardiovascular to environmental exposures,

to genetic and psychiatric, many different traits.

And then within those different trait groups, we were able to find studies. We were able to basically scrape the statistics on those studies,

to look at the cohort sizes, the impact factor of the studies,

and then rank the probes according to their statistical significance.

So we were able to prioritize the studies and get a broad diversity of trait groups,

and prioritize the probe selection that way. And then from all of our other studies,

we were able to just downsample and select a small number, from what we deemed as the lesser,

statistically robust studies, but hopefully still capture some markers,

from the traits that were examined, in those in those studies. So we had roughly 20 major trade groups

with thousands of studies and thousands of some traits, which we think encapsulate a lot of different biology.

The next phase of the design was in recent years, there’s been a lot of large scale, whole genome

bisulfide sequencing projects. So we wanted to sort of combine these advantages.

We discussed for the microarrays, where you could have a select CPG site with high quantitative precision,

with the knowledge that’s been, gleaned from these studies.

So we looked at, thousands of different single cell and sorted

GBS data sets, and we had different target biology groups that we wanted to find markers for.

So we were particularly interested in tissue specific and cell specific CPG sites.

So these are markers in the genome CPGs where the methylation

or the absence of methylation essentially defines the cell type and distinguishes it from all the other cell types

that might be in the tissue, or in the body. So we are very interested in those.

And we did a lot of analyzes over thousands of methyl omes to find markers that distinguish hundreds of different cell types.

And I’ll talk a little bit more later on about some of the applications of that.

We were also interested in sites that might where the methylation might actually be correlated with gene expression.

So in these studies, you do any EWAS, if you identify a site that associates with a trait, you ultimately

want to understand what it might be doing or what the biology is behind that.

Methylation trade association. And usually if,

if a methylated site is going to have an impact on a phenotype, it’s probably going to be impacting gene expression or genome stability,

in one way or another. So we were able to leverage these matched,

GBS, RNA-Seq data sets to find the CPGs sites where the methylation actually correlates with the gene expression.

So that’s another example of one of the design groups that we targeted. Then we had about 15 more other biology groups

that we were able to mine GBS and other sources for—

but altogether there’s a lot of new biology, a lot of the new probe sites target, Cis-regulatory elements.

So these are functional elements in the genome that are predicted to have a role in gene expression and regulation of the genome.

So there’s about over 100,000 new probes, sites that target these different biologies,

that might have a lot of discovery power and might lead to novel associations.

And not just for methylation, but hydroxymethylation as well, which is a related epigenetic mark.

So yeah, it was leveraging both prior studies as well as new, whole genome bisulfide sequencing

data, where we were able to identify a lot of these markers. Awesome, it seems what I would say, David, like it’s, super comprehensive.

Right? So, I’m excited to learn a little bit more about how you all were able

to pick the the probes in the content design of that MSA array. Now, just very briefly as well,

if you were to pick like let’s because there’s so many let’s, let’s confine it to like one or two areas or pieces of content.

What do you think would be most useful for those performing those EWAS studies?

Does anything stick out in particular? Yeah, well it’s hard.

I think they’re all going to be pretty useful. Maybe I’m a little biased, but, I really am excited about the cell specific methylation

patterns, that we were able to identify and validate, in some of our,

in some of our initial studies. I think that most diseases and traits

ultimately manifest as altered cell type proportions,

or altered functioning of the tissue type or the trait manifests. And these markers will really enable high

resolution deconvolution. Or I think they’ll enable the study of more rare

cell types in heterogeneous bulk samples. So normally when you do an egg wash you take a biopsy of tissue.

But that tissue has many different cell types. And if the one that’s implicated in the trait is relatively rare,

then you need more markers for that cell type. But I think as a result of this cell

specific methylation, it’ll enable it’ll give the ability to really deep convolve bulk tissues and understand

how different traits are altering the more rare subtypes. And to that point, previous studies have tried and have,

have done deconvolution, but it’s on a more coarse level. So for example, in the brain,

I do some work on neurodegeneration. Oftentimes the deconvolution will only be looking

at neurons and non-neurons. But now as a result of these single cell,

GBF studies that have been done in the brain in the last couple of years, we know that there are hundreds of cell types.

And now I think we can actually develop, references and deconvolution panels

to better estimate the proportions of more rare cell types. In EWAS studies,

both to covary for cell types, but as well to identify the populations that might be vulnerable

in a given exposure disease. So I think that’ll be one of the major strengths of this new platform.

Yeah. And that one sounds sounds exciting for sure. Well, we’re getting to the end of this, this podcast.

Kind of one more question for for each of you both. Nicole, I don’t mean to to leave you hanging, but I’m going to save yours for the end.

So I’m going to, bombard David with with one last question here.

Before we we start to close this out, so just kind of again, adding on to the topic

we’re already mentioning regarding, you know, how we created the MSA, you know, the content that you put on it, what you think would be most useful for,

Future Directions in DNA Methylation Research

for researchers, especially those with EWAS kind of the single cell profiling, you know, you and Wanding where we’re really heavily involved

in just the overall creation of this MSA beadchip, I think that’s super clear. You published a preprint as well.

Maybe by the time this is published, it’ll be pushed to publication regardless. Well link it in the show notes, but,

I know a big concern of not only, you know, users of the microarrays, but even down to whoever’s receiving

the end result from the micro array, whether it be a diagnostic tool or a report, is always wanting to know about the technical performance

and the biological validation, which, you know, it was run on. Can you just talk very briefly about that?

David. Yeah. So basically we go through a series of,

sort of technical experiments to check the performance. So we’ll generate DNA and methylation profiles for cell lines,

and we’ll look at the cross replicate correlations. And in those experiments,

the MSA readings between technical replicates were highly correlated

and performed at a similar level to previous platforms. We also compared the,

accuracy, comparing it to other sequencing technologies, and found extremely high correlations,

with, both, GBS sequencing as well as from other capture

sequencing approaches. So the performance is,

is as good as the previous platforms in our experiments, which we were really happy and relieved to see.

In terms of the biological variation or validation, we did we’ve now profiled.

So the original preprint, doesn’t have all of our data, but,

our revisions. Well, we’ve profiled hundreds of tissues,

both for methylation and hydroxymethylation. And our, across dozens of tissue types.

And we’re seeing very nice capacity to do deconvolution of the tissues.

When we do unsupervised clustering, we’re able to see that the array is, the methylation profiles are clearly,

clustering the tissue types. And we validated a lot of our markers, which is great.

And then we’ve also tested the epigenetic clocks that we included in the design,

some of the major ones that are used in the fields. And we see very high correlations between the predicted epigenetic ages

as well as the with the chronological ages of our samples.

So and there’s a lot more biology, that will be in our paper. But so far the technical performance looks great.

And we’re learning a lot of interesting biology from the samples, the actual, tissue samples that we profiled.

So… -Great. Yeah, I’m excited to read more about the, the, biology involved in that validation.

Once we get those those edits to the pre-print, and as well and it’s published so sounds great.

Back to you, Nicole. Sorry to leave you hang in there for a second. You know, last question for you.

What excites you most about the methylation screening there? Right? Like, why do people have to be using this?

Yeah. So I’m I’m really excited about the methylation screening array because we’re really fulfilling

a very large user need. And like I mentioned before, I think this tool will enable

the next generation of methylation research. So really the application of methylation

profiling in very large screening projects. And we might even see this at the,

you know, government initiatives level. For example, you know, Million Veteran Program.

There’s a lot of there’s a lot of new government initiatives and biobank projects that are seeking to be able

to better predict diseases and to better predict how to treat diseases and prevent diseases.

So I think, getting methylation, all of the methylation insights that have been found, you know, we in the epigenetics community

Closing Remarks

know how powerful DNA methylation is as a biomarker for human disease.

But getting those insights to scale and actually impact,

human health in a very large way is really exciting to me. And I’m just, you know, I have kind of the

what I enjoy most about my job is just seeing all the applications that have been developed and continued to be developed.

So I think we’re going to see more aging clocks and more methylation

risk scores and new classifiers, and all sorts of common disease areas

that will be enabled by this methylation screening array. Yeah, definitely. I okay, I have to sneak one.

One last question for you, Nicole too, because I’m just personally, selfishly interested.

What’s next in the methylation products in terms of development? So just very quickly, is there anything you all are looking at in the future?

I know you mentioned, you know, maybe reevaluating or relooking at the,

whole genome bisulfate sequencing and maybe, you know, adding better tools for more of that discovery based purpose.

But like, is there anything on the pipeline? What can you tell us about that? Yeah, definitely.

I’ll say broadly, Illumina is investing in DNA methylation technology, whether it’s microarrays or sequencing,

based approaches. And first, to just dispel the myths, I think, that have been present

for over a decade now is, you know, people have been forecasting the death of microarrays.

And what we’ve seen in the methylation microarray business is that,

you know, consistent growth over year over year. And,

as a result, Illumina is actually investing more into methylation microarrays then than, they ever have been.

And I can’t share details of specific products.

But, what I can say, yes, we are interested in in different types of chemistries that are emerging.

And, but really our goal is to just amplify a lot of the benefits that bring people to the platform.

So like we’ve mentioned, reproducibility, scalability, affordability, we’re working on all of those things.

And then we’re also going to be focusing a lot on the user experience. And like David said there’s a

there’s a great set of existing tools. But how can we make it easier for, say,

new researchers to get from data to biological insights faster? And,

so and, you know, I can promise you that at the forefront of all our product developments will be,

really trying to understand what our users need in terms of new content. And, our, my most important priorities is staying really close

to the user community. I always try to make myself available so you can you can reach out to me

if you have any product ideas or, ideas to improve, say, the content of our products.

I can be reached on X at Nicole Renke and or on LinkedIn.

Also, Nicole Renke and I’m always really interested in connecting with our users.

Yeah, I love that. That’s awesome. Yeah. The microarray, it’s not going anywhere. It’s here for, a very, very long time.

Thanks to both of you. You know, Nicole, you just gave your information in terms of, you know, where where listeners can come to find you.

David, what about yourself? If anyone wants to reach out, I’m sure you have your your UPenn profile and whatnot.

But how can people get ahold of you if they’re interested— or look at some of your work? Yeah.

My email in the lab. So I’m in the lab of Dr. Wanding Zhou. Which, hopefully we can drop a link to the lab website, but,

we’re doing some really cool stuff with DNA methylation and developing informatics techniques, so, I can be reached through there as well.

And hopefully, yeah, hopefully someone reaches out. Of course. Alright. My fun question at the end of the podcast.

I was itching to get to this one. I don’t care who goes first. It’s if you could be any animal in the world, what would you be and why?

You have to give an answer. Okay,

Well, for me, I think I would be a sea turtle because I could just move

so gracefully through the water and they just look so calm. I think they’re just a beautiful creature.

Love that. I don’t think I’ve ever gotten that answer before. So that’s a good one. David. What are you thinking?

Well, I definitely don’t want to be a mouse. I don’t want to be an I don’t want my methylation…

But, yeah, maybe a bird of some sort. You know, they have the ability to fly and go long distances, see places.

So we’ll go with an eagle maybe. There we go. That’s a good one. But no, this has been great.

Nicole and David, thank you so, so, so much. So excited to learn about the new methylation screening array and the history behind the microarrays within Illumina.

Thanks everyone else for listening. And joining us at the Everything Epigenetics podcast. Remember, you have control over your epigenetics.

So tune in next time to learn more about how. Thank you all. Thank you Hannah, take care. -Thank you.

Methylation Screening Array: A Revolutionary Tool in Epigenetics

About this Guest

Welcome and Introduction

History of Methylation Microarrays

Applications of EWAS in Research

Unveiling the MSA Array: Vision and Design

Technical Validation and Biological Insights

Future Directions in DNA Methylation Research

Closing Remarks

More Episodes

Breaking Down Epigenetics: Sperm, Seminal Plasma, and Generational Impact

Methylation Screening Array: A Revolutionary Tool in Epigenetics

Heart Health and the Future of Epigenetics