A large cross-population GWAS meta-analysis (168,007 AF cases) integrated with proteomic data identifies hundreds of AF loci, implicates cardiac and TGF-β pathways, finds causal risk factors and proteins via Mendelian randomization, and shows improved prediction when combining polygenic and protein scores.
0:00Welcome to Base by Base, the papercast that brings genomics to you wherever you are. Thanks for listening, and don't forget to follow and rate us in your podcast app. So I want you to picture a healthy human heart.
0:11It works like this perfectly synchronized orchestra. Yeah, exactly. It's a really tight rhythm section Right. But then for, I think the number is nearly 59.70000000 people worldwide. That electrical rhythm just, well, it breaks down into total chaos.
0:27It does, yeah. And that's the condition we call atral fibrillation. Exactly. So instead of a conductor keeping time, it's like every single musician suddenly starts playing at their own tempo. Which, musically speaking, sounds awful.
0:37Right. And medically speaking, it's even worse. At least to palpitations, a massive drop in exercise capacity, and some really serious cardiovascular complications. Yeah, strokes, heart failure. It's incredibly dangerous.
0:49But and this is the crazy part. Imagine a world where we could predict this chaotic rhythm years before it actually happens. Just by reading a person's genetic blueprint. Yeah. And looking at the proteins floating around in their blood.
1:03So the question we're asking for today's deep dive is how could expanding our genetic mapping, beyond just one ancestry, fundamentally rewrite our understanding of how and why the heart loses its beat?
1:15Well, today we celebrate the work of Shuai Yuan, Chi Chen, Susanna C. Larson, Scott M. Damrauer, and a massive collaborative research team from institutions like the University of Pennsylvania and Karolinska Institute, who have advanced our understanding of the genetic etiology of atrial fibrillation.
1:32It's just a huge monumental effort. Massive. We're basing this deep dive on their article, cross-population GWAs and proteomics improve risk prediction and reveal mechanisms and atrial fibrillation, which is published in the journal, Nature Communications on July 11, 2025.
1:48Such an incredible paper. It really is. But to truly appreciate this breakthrough, we 1st need to look at why previous genetic maps of the heart were, you know, incomplete. Right, the historical gap. Exactly.
1:59So we have to look at the scale of this epidemic. The global population is aging, right? Yeah, definitely. People are living much longer. And because of that, atrial fibrillation or AF is just surging, especially in places like North America and Europe.
2:12Right. And obviously, lifestyle and environment play a huge part of that. They do, yeah. Diet, stress, all of that. But the underlying genetic susceptibility to AF is substantial, like the physical architecture of the heart, how it handles electrical impulses, how it responds to stress.
2:31Much of that is coded right there in our DNA. Which is why researchers have been doing these genome wide association studies or GWS for years, right? Trying to map out those exact genetic coordinates. Exactly.
2:43And prior GWS actually did uncover over a 100 risk LOSI for AF. Okay, wait, let's unpack this for a 2nd because almost all of those previous studies only looked at European populations. Yeah, that's the big blind spot.
2:56I mean, isn't that like trying to solve a giant jigsaw puzzle when you've only been given pieces from like the top left corner? That's a spot on analogy. The lack of ancestral diversity is a massive hurdle in genomic.
3:07Because you're just missing so much human data. Exactly. When you only look at one population, you miss variants that might have smaller effect sizes, you just lack the global statistical power. Right. You also miss population specific genetic low sci.
3:22But perhaps most critically, you struggle to isolate the shared genetic low side. You mean the universal biological mechanisms. Yeah, the ones that drive the disease across all humans, regardless of their ancestry, because if you can find those universal mechanisms, you can develop universal therapies.
3:38So to build a more comprehensive map and fix that puzzle, this team went huge. Oh, unprecedented scale. They did a cross-population GWS men analysis pulling data from over 2000000 individuals. Two million.
3:52That is just staggering. It is. It included 168,7 patients with AF and nearly 2000000 controls without it. And they finally spanned different ancestries, right? Yes. European, East Asian, African and ad-mixed American ancestries.
4:06Which is fantastic. But the scale of the data is just the start, right? Finding a genetic signal, like a specific coordinate on the DNA that correlates with AF, is only step one. Right, finding the locust.
4:17Wait, let's explore that for a 2nd for you listening. Because if they already have 2000000 people's genomes and they spot a region of DNA, highly linked to AF, why do they need complex computer programs to figure out the gene?
4:29The data is right there. Can't they just look at the genes sitting at that spot? You would think so, but the human genome is incredibly complex. A genetic variant associated with a disease might sit in this vast genetic desert far away from any known gene.
4:45Or it might sit right between 2 completely different genes. It might not even affect the gene closest to it. It could be a regulatory switch that, um, folds over in three-dimensional space to control a gene 1000000s of base pairs away.
4:59Ah, okay. So finding the locust is like identifying a suspicious neighborhood, but it doesn't tell you which house the problem is coming from or who actually lives there. Exactly. So the researchers deployed this really comprehensive gene prioritization framework.
5:12Meaning they used multiple tools to narrow it down. Yeah, 6 different computational methods. Rather than just guessing based on proximity, they use tools like May GMA and EQTL colloquialization. What does EQTL colloquialization actually do?
5:26Well, it essentially looks at actual human heart tissue to see if the genetic variant in question directly alters how much of a specific protein is being produced. Oh, so they are tracing the wire from the genetic switch directly to the light bulb inside the heart tissue.
5:41Spot on, and that rigor is what gives us the headline results here. They identified 525 genetic loci reaching genome wide significance. And of those, 379 were completely novel. They had never been reported before.
5:54That fundamentally expands what we know about the biology of the heart. But when they looked across the different ancestries, did any specific genes stand out as universal? They did. They found 2 specific genes, PITX2 and ZFHX3, that were shared risk losi across all 4 ancestries.
6:12What do those genes actually do in a healthy heart? So PITX 2 is deeply involved in structural development and how the heart handles calcium ions. And calcium is the messenger for contraction, right? Exactly.
6:23So if PITX 2 is dysregulated, calcium handling gets erratic, it leads to rogue electrical sparks that trigger the chaotic rhythm. And ZFHX 3 is similarly vital. Losing its function leads directly to atral dysfunction and physical remodeling of the heart tissue.
6:40The electrical grid gets structurally compromised. Now, I do want to mention the breakdown of those 525 lowsi across the populations because it highlights an ongoing issue. Right. The numbers reflect the available sample sizes.
6:51Yeah. So it was 483 LOSI in Europeans. 29 and East Asians, 5 and Africans, and 2 in admixed Americans. Which just shows how much discovery potential is still out there for non-European populations. Absolutely.
7:04Okay, so having this genetic map is incredible, but genetics is only half the battle. If we want to prevent AF, we need to know what real world habits are waking up these dormant genetic risks. Right, the lifestyle factors.
7:16And to figure that out, the team used Mendelian randomization. Yes. Let me actively jump in and clarify Mendelian randomization or MR for you listening, because it's brilliant. So instead of relying on observational studies, which can be super misleading, Mendelian randomization acts like nature's own randomized controlled trial.
7:35It's such a powerful tool. It really is. Because your genes are locked into birth. Researchers can see if a genetic predisposition to a trait, like high cholesterol, actually causes the disease, rather than just hanging around the scene of the crime.
7:49Right. It completely bypasses the environmental confounders that plague normal observational studies. So when they ran those modifiable risk factors through this MR analysis, what actually caused the fibrillation?
8:01They confirmed definitive causal links for traits like childhood and adult BMI, waste to hip ratio, smoking, alcohol consumption, and type 2 diabetes. Which, you know, physical damage from obesity or smoking makes intuitive scent.
8:15drives inflammation, yeah. But they also found insomnia was a causal driver. Wait, insomnia. Really? How does a lack of sleep fundamentally rewire the electrical grid of the atria? It comes down to the autonomic nervous system.
8:27When you suffer from chronic insomnia. Your body is subjected to sustained physiological stress. Like being stuck in photo flight mode. Exactly. Your sympathetic nervous system becomes overactive. This floods the heart with the adrenaline over long periods.
8:42So the heart is constantly being told to race. Yes. And that chronic overstimulation leads to increased oxidative stress and inflammation, which eventually causes microscarring or fibrosis in the delicate atrial tissue.
8:55And then the electoral signals hit that scar tissue fragment and start circling back on themselves. And that is what creates the fibrillation. That is fascinating how a behavioral state like wakefulness translates into literal physical scar tissue.
9:09It's incredible biology. It is. But this brings us to the twist in the proteomics data. We mentioned at the very beginning of this deep dive that there was a finding regarding a circulating protein that completely upends a long-standing clinical assumption.
9:23Oh, yes. the NT ProBNP finding. Tell me about this because this protein is a huge deal in cardiology. It is. In traditional observational settings, NT ProBNP is a gold standard biomarker. When a patient comes into the hospital suffering from severe atrial relation, their blood levels of this protein are exceptionally high.
9:45Right. And if you always see high levels of a protein during a cardiac crisis, the logical assumption is that the protein is contributing to the pathology. You'd assume it's part of the disease process, yeah.
9:56But when the researchers applied Mendelian randomization to determine the actual causal direction. They found the exact opposite. They discovered that a genetic liability to AF is actually associated with reduced levels of NT ProBNP.
10:11Yes. That feels completely contradictory. If it's a marker of heart failure, shouldn't genetic risk for the disease elevated? It totally seems like a paradox, but it actually reveals the beautiful compensatory mechanics of the human body.
10:23Because MR shows us the true causal direction. This inverse relationship proves that the high NT pro BMP levels we see in patients aren't the cause of the Fibrillation. They are the consequence. Exactly.
10:35When the heart's rhythm breaks down. The atrial muscle stretches and experiences immense hemodynamic stress. In response to that trauma, the heart pumps out NT, pro B, and P, to try and reduce blood pressure, and decrease the workload.
10:49So it's a protective secondary stress response. It's the fire alarm, not the arsonist. That's a perfect way to put it. Having genetically lower baseline levels of this protective protein might actually be part of what leaves your heart vulnerable to the chaos in the 1st place.
11:05It perfectly illustrates the danger of relying solely on observational data. We see the firefighters at the burning building, and assume they started the fire, just because they're always there. But the genetics prove they're just responding to the emergency.
11:17And this distinction fundamentally transforms how we think about risk prediction, because the researchers didn't stop at discovery. They wanted to build something doctors could actually use. Right. They integrated their genomic data with proteomic profiling, analyzing 87 circulating proteins, to build incredibly powerful risk models.
11:36It's like a supercharged crystal ball. Let's highlight the performance of combining the polygenic risk score, or PGS, with a protein score, the pro S. Can you break down the stats for the listener? Sure.
11:48So the combined model achieved an AUC of .823. Okay, for anyone who doesn't live and breathes statistics, hitting an AUC of .823 for a highly complex multifactorial disease like AF, is just a massive leap forward.
12:03It really is. Usually polygenic risk scores for complex traits hover in the .6 to .7 range. So .823 means the model is highly capable of identifying those at extreme risk. Individuals in the top 10th percentile of this model had a more than sixfold increased risk of developing AF compared to the bottom percentile.
12:22That's incredible. It allows doctors to predict a patient's risk of AF with high accuracy decades before it happens. Exactly. So how does this actually shave clinical practice? Well, targeted AF prevention can now focus with certainty on weight management, lowering blood pressure, and behavioral intervention.
12:38Like reducing screen time for insomnia or cutting alcohol. Yes. Furthermore, identifying pathways like TGF beta signaling opens the door for novel drug developments. TGF beta is heavily involved in cardiac fibrosis, right?
12:52It is. It dries the scarring. So if pharmaceutical companies can target that specific pathway, we might be able to halt the structural damage before the electrical circuits ever miss fire. We could treat the architecture of the heart before the rhythm breaks down.
13:06Though we do have to note the study's limitations. Despite the cross population effort, the statistical power in non-European ancestries was still limited. Right. Due to those smaller sample sizes, we mentioned earlier.
13:19It proves we desperately need more equitable genetic databases. We absolutely do. And measuring proteins in the blood while minimally invasive and super convenient might not perfectly capture what is happening directly in the localized heart tissue.
13:32Because the blood is picking up signals from the whole body. But even with those limitations. The central insight here is just profound. By uniting multi-ancestry genomic data with circulating protein analysis.
13:43This research expands the genetic map of atrial fibrillation with 100s of new risk glossi. And it cleanly separates causal drivers from biological bystanders. Yes. This multi-inzymetic approach fundamentally enhances our ability to predict, prevent, and eventually treat the world's most common cardiac arrysmia, which leaves you, the listener, with a final question to ponder.
14:06Go for it. What does this mean for the future of cardiovascular care? When a simple blood draw and a DNA swab could map out your heart's rhythm decades before it ever misses a beat? It's a whole new frontier for medicine.
14:17This episode was based on an open access article under the CCBY 4.0 license. You can find a direct link to the paper and the license in our episode description. If you enjoyed this, follow or subscribe in your podcast app and leave a 5 star rating.
14:32If you'd like to support our work, use the donation link in the description. Now stay with us for an original track created, especially for this episode, and inspired by the article you've just heard about.
14:42Thanks for listening, and join us next time as we explore more science base by base.
