0:00Welcome to Base by Base, the paper cast that brings genomics to you wherever you are. Thanks for listening, and don't forget to follow and rate us in your podcast app. Hmm. So imagine you're in this vast dark utility room that controls the body's entire immune system.
0:14and you're trying to figure out the wiring. For decades, when we've tried to find the genetic causes of complex diseases, like lupus or rheumatoid arthritis, we've only been able to explain maybe a 3rd of the circuitry.
0:27Right, about a third, it's been a persistent problem. We use these incredible tools, genome wide association studies, or GWAS, which are fantastic at finding the exact, the addresses of genetic risk variants.
0:41They give you the location on the map, a pinprick on the DNA sequence that's linked to a disease. Exactly. But most of the time we find the address, we flip the switch, and the lights just don't come on.
0:50We can't figure out the functional mechanism. It's what people call genetic dark matter. Yeah, that's it. We know the location, but we have no idea what the job is. So the big question is, why do we keep missing that functional link between a tiny genetic change and the, you know, the huge disease it causes?
1:08Well, that's what's so exciting about the paper we're diving into today. It feels like a total game changer. The study we're looking at, it seems to have cracked the code for about 50% more of these mysterious genetic locations.
1:19It really is a huge leap. That inability to get from a genetic location to its actual function has been the biggest bottleneck in translating all this genomic data into, well, into therapies. So this work is a massive step forward.
1:34A massive step. And we really have to acknowledge the scale of the effort here. Today, we're celebrating the work of Zipping Mu, Luis Barrero, Yang Li, and their teams. Spanning institutions like the University of Chicago and the CZ Biohub Chicago.
1:47Absolutely. Their focus on the functional interpretation of genetic risk is, well, it's exactly what the field has been needing. Okay, so let's unpack that core problem they were tackling. For years, the gold standard for finding the function of a risk variant was to look for something called an EQTL.
2:05An expression quantitative trait locus. Right. And an EQTL is basically measuring the final result, right? The light bulb itself. That's a great analogy. It links a genetic variant to a measurable change in gene expression.
2:18So how much RNA the cell is actually producing. And if your genetic risk factor causes G X to produce way too much product, then you've probably found your culprit. That's the idea, but as you said, EQTLs alone only account for about a 3rd of the known GW's hits for immune traits.
2:35Which tells us that the other 2 thirds aren't changing that final output, at least not in a way we can easily measure in a standard resting sample. Exactly. So they must be working one step earlier, not at the light bulb, but at the regulatory switches themselves.
2:47And this brings us to chromatin accessibility, the KQTLs. Precisely. Chromatin accessibility, quantitative trait, Losi. So if EQTLs are the light, KQTLs are literally the switch on the wall. You've hit it right on the head.
2:59They link a genetic variant, not to the amount of RNA, but to changes in chromatin accessibility. How open or closed the DNA is. Exactly. Chromatin is the structure that packages your DNA. If a region of chromatin is open, it means regulatory elements, things like promoters and enhancers can physically get in there and bind the DNA.
3:21Turning the switch on. Turning the switch on. So the hypothesis was that many of these unexplained variants work by changing this physical accessibility, not necessarily the immediate RNA output that you can measure downstream.
3:33Okay, so the genetic change is flipping a physical switch, but the light, the detectable gene expression, it might only come on under very specific conditions. Conditions that standard mapping often misses.
3:46Which is a huge shift in focus. And to test that, you need a methodology that's just as dynamic. And that's what they did. They didn't just look at a bunch of cells in a tube. They created a single cell, chromin accessibility map scat tack from, and this is amazing, roughly 280,000 individual cells.
4:02280,000 single cells. That number is just huge for this kind of analysis. It is. And it was from 48 different individuals. But what really elevates this study, you know, is the context. They didn't just study healthy people.
4:16Right key innovation. The cohort wasn't static. It was disease relevant. It included 20 individuals with active COVID-19 and 19 donors who were convalescing. That clinical context, that urgency allowed them to capture these genetic effects during a real, high stakes immune response.
4:34I find that so insightful. But how do you handle that complexity? You're not just dealing with 7 broad immune cell types anymore. You have cell states that are changing, I mean, by the hour. Well, they move past simple static categories.
4:47Instead of just putting cells into discrete boxes like TCell or BCell, they use a computational approach called topic modeling. Topic modeling, like for text analysis. The very same. It allowed them to map continuous cellular states and trajectories.
5:00So the actual path the cell takes as it responds to an infection. So you can see the journey, not just the start and endpoints. Precisely. And that was absolutely necessary to find what they call dynamic KQTLs.
5:15These are the regulatory effects that only pop up when a cell is transitioning from one state to another. When it's in the middle of a fight. Exactly. They were looking for genetic effects that only show up when the cell is actually busy working.
5:26Okay, let's get to the results because the numbers here are just, they're staggering. We really are. So first, the sheer explanatory power. By integrating these accessibility changes, the KQTLs with the expression changes, the EQTLs, they just, they dramatically reduce that genetic dark matter.
5:45The total percentage of GW is low-si, they could functionally explain, jumped from, what was it, about 33%? 32.9%, yeah. Right. From about a 3rd to nearly 60%, 57.3% to be exact. It's like someone turned the lights on for over half of these risk locations overnight.
6:02It really validates the whole idea that this regulatory variation, the switching mechanism, is the primary driver for a lot of these complex immune diseases. But here's where it gets a little tricky. The paradox. The paradox.
6:14is where it gets really interesting for interpretation. The sharing versus specificity problem. So what you found is that the physical switch, the act of opening the chromatin, is often broadly shared, right?
6:27Very broadly shared. Less than 20% of all the KQTLs they found were restricted to just a single cell type. So the genetic variant seems to open that bit of chromatin in multiple different immune cells.
6:38Okay, so the blueprint for opening the switch is widespread. But, and this is the key point, the light bulb, the resulting change in gene expression, is extremely restricted. Wait, hold on. So the switch is physically accessible in, say, 80% of cell types.
6:53But the gene only gets turned on in one of them. What's acting as the final gatekeeper there? That restriction confirms that the cell state itself is the most critical filter. The study showed that when a KQTL did cause an EQTL when that light actually came on, 84.5% of those pairs were found in only a single specific cellular context.
7:13So, the genetic vulnerability at the switch is broad, but the functional consequence, the cell actually reading that accessible DNA is incredibly narrow. Which tells us immediately that you can't study these things in a vacuum.
7:27The cell has to be in the right physiological moment for the genetic effect to actually matter. You got it. And that's where their dynamic modeling really paid off, especially with the COVID-19 data. Yeah, tell us about that.
7:37How did that dynamic context reveal the functional relevance? So the topic modeling identified a specific cell state they called K17. It's a continuous path that mostly involves these highly active CD 8 effector memory T cells.
7:52The frontline soldiers in the immune fight. Exactly. And this state was highly enriched in the COVID-19 patients, but here's the kicker. They then showed that the genetic regions associated with this K 17 state were significantly enriched for heritability linked to severe hospitalized COVID-19.
8:07So they connected a specific active cell state directly to the genetic risk for a severe outcome of the disease. And by mapping that continuous trajectory, they found over 4200 dynamic accessibility changes.
8:20changes that only appeared as the cell was moving into that activated state. That's incredibly compelling. And it applies to other diseases too, right? I think they mentioned that for rheumatoid arthritis, over a 3rd of the risk low side, that colloquialized with an accessibility change, show these dynamic effects.
8:38That's right. It suggests the genetic risk for Omar Array might not be about what your immune cells look like when they're resting, but what happens to them the moment they get activated. So we have this massive influx of new knowledge, but it also creates a new problem, right?
8:53The interpretation challenge. Yes. The biggest group of these newly explained locations are what the authors call KQTL only. They clearly affect accessibility, the switch is flipped, but the light isn't on yet.
9:05There's no detectable change in gene expression. And that is the big question mark hanging over the functional utility of all this new data. If you have a press switch but no light, it's really hard to confidently say you found the causal gene.
9:17Or even the right context, are we just missing the cell type where the expression happens, or are we missing the exact stimulus that triggers it? That's the $1000000 question? So if KQTLs alone aren't quite enough, what's the new gold standard?
9:32How do we get to that high level of confidence to say, this is the causal gene? The most robust evidence, the highest confidence. It only emerges when you see a triple convergence. You need the GW arts risk location.
9:44Okay, the address. The KQTL signal, so the accessibility change. and the EQTL signal, the expression change. All aligning perfectly in the very same cellular context. You need all 3 for that mechanistic clarity.
9:57And the inflammatory bowel disease case study really showed this in action. It's a beautiful example. They tracked a known IBD GWS, locus. And the evidence didn't fully converge until they looked specifically at monocytes.
10:11What did they find there? The chromatin accessibility pick was specific to monocytes, and the resulting expression change for the target gene, Kari D9, was also specific to monocytes. And didn't they even show a physical connection too?
10:23Yes, that was the final piece. They used a model called Activity by Contact, or ABC, which predicts physical enhancer to gene links, and that link to the car D9 promoter was only present in monocytes. So genetic location, accessibility, expression, and the physical wiring all pointed to one gene, card D9 in one cell type, monocytes.
10:46That is the new gold standard for causal gene identification. And that really drives home the core lesson of this paper. That widespread sharing of accessibility changes across all these immune cells is functionally ambiguous.
10:58Exactly. Just mapping KQTLs isn't enough anymore. We have to map these genetic effects in highly specific disease relevant and dynamic cell states to get past that ambiguity. Of course, they did note a limitation, right?
11:10The sample size. A fair limitation. 48 donors provided fantastic context, but that sample size still limits the statistical power to definitively find these effects in rarer but maybe really crucial immune cell subtypes.
11:23So this work really sets the stage for future, even larger population scale study. It does, and it demands that those future studies also capture the single cell dynamic information. So here's the central insight we want to leave you with.
11:34Genetic variants that affect chromat and accessibility. These KQTLs have unlocked a huge amount of previously unexplained genetic risk in immune diseases. We can finally see the switches we couldn't before.
11:47However, translating that knowledge into actual function requires moving beyond these static snapshots. The true causal gene and the true causal cell type are only confidently identified when that accessibility signal and the expression signal aligned perfectly in a specific functional cellular context.
12:05All about context. It is. And this shift from viewing our immune systems as a static circuit board to a dynamic, actively switching machine well. Makes you wonder, what does this intense focus on dynamic regulatory variants mean for developing cell state-specific therapeutics in the future?
12:21This 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.
12:36If 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.
12:45Thanks for listening and join us next time as we explore more science base by base.
