Disease emerges from disruption of networks, not failure of single genes
In the long human struggle to decode the origins of neurodegeneration, UCLA researchers have taken a meaningful step forward — not by finding a single answer, but by revealing that the question itself must be asked differently. Using high-throughput genetic screening methods, the team identified new risk genes for Alzheimer's disease and progressive supranuclear palsy, and found that these genes do not act alone, but as members of disrupted biological networks. Published in Science, the work reframes how medicine might one day approach treatment — not by silencing a single faulty gene, but by restoring the harmony of an entire regulatory system.
- For decades, genome-wide studies have pointed to disease-linked regions of DNA without being able to name the specific variants actually driving harm — a bottleneck that has stalled therapeutic progress.
- The UCLA team broke through by testing 5,706 genetic variants simultaneously using massively parallel reporter assays, identifying 320 high-confidence functional variants far faster than traditional one-at-a-time methods would allow.
- A second layer of validation using pooled CRISPR screening confirmed the findings, lending unusual methodological rigor to results that named entirely new risk genes — C4A, PVRL2, and APOC1 for Alzheimer's, and PLEKHM1 and KANSL1 for PSP.
- The deeper discovery was architectural: in progressive supranuclear palsy, multiple scattered variants appear to pile onto one another, collectively dismantling the transcription factor networks that regulate gene activity in specific brain cell types.
- The field is now pointed toward network-based drug strategies — therapies that address coordinated biological disruption rather than chasing single proteins — though translating this framework into clinical treatments remains the next hard frontier.
Researchers at UCLA have opened a new corridor in the search for the genetic roots of Alzheimer's disease and progressive supranuclear palsy, deploying high-throughput laboratory techniques to identify multiple previously unknown risk genes for both conditions. The work, published in Science and led by Dr. Dan Geschwind and first author Yonatan Cooper, also advances a broader argument: that neurodegeneration is less a story of isolated broken genes than of disrupted biological networks.
The problem the team set out to solve is one of genetics' most persistent frustrations. Genome-wide association studies can identify chromosomal regions linked to disease, but those regions contain dozens or hundreds of genetic markers traveling together — and distinguishing the truly harmful variants from passive bystanders has long resisted easy solution. To cut through this noise, the team used massively parallel reporter assays to screen 5,706 variants across 34 disease-associated regions at once, yielding 320 functional variants identified with high confidence. They then validated a subset of 42 using pooled CRISPR screening across multiple cell types, adding an independent layer of confirmation.
The concrete findings were significant. Three new Alzheimer's risk genes — C4A, PVRL2, and APOC1 — were implicated, along with PLEKHM1 and KANSL1 for progressive supranuclear palsy. But the more conceptually striking result was the evidence that in PSP, genetic variants scattered across the genome act additively, collectively disrupting the transcription factors that switch genes on and off in particular brain cell types. Disease risk, the data suggested, emerges from the cumulative unraveling of regulatory networks rather than from any single point of failure.
Geschwind was careful to frame the high-throughput approach as a bridge rather than a replacement — a way to connect population-level genetic signals to actual biological mechanisms. The therapeutic horizon this opens is one where targeting a network of genes simultaneously may prove more effective than conventional single-protein drug strategies. The next phase of work will test whether the newly identified genes interact in ways that can be therapeutically exploited — a question whose answer will determine how quickly this scientific advance reaches patients.
Researchers at UCLA have mapped a new path through one of genetics' thorniest problems: finding the actual culprits among thousands of genetic suspects. Using a combination of high-throughput laboratory techniques, they identified multiple previously unknown risk genes for Alzheimer's disease and progressive supranuclear palsy, a rarer neurological disorder with similar underlying pathology. The work, published in Science, suggests a fundamentally different way to think about how genes contribute to neurodegeneration—not as isolated actors, but as members of coordinated networks.
The challenge that prompted this research is deceptively simple to state and fiendishly difficult to solve. When scientists conduct genome-wide association studies, they scan the genetic code of large populations to find regions linked to disease. But each region contains dozens, sometimes hundreds or thousands of genetic markers that travel together through families and populations. Researchers know the region matters, but they cannot easily tell which specific variants are actually doing the damage and which are just along for the ride. Identifying the functional variants—the ones that genuinely alter disease risk—has remained a major bottleneck in modern genetics.
The UCLA team, led by Dr. Dan Geschwind and first author Yonatan Cooper, deployed a technique called massively parallel reporter assays to test 5,706 genetic variants across 34 disease-associated regions simultaneously. Rather than testing variants one at a time, this approach screens thousands in parallel, dramatically accelerating the process. From that initial screen, they identified 320 variants they could confidently classify as functional. To strengthen their findings, they then used pooled CRISPR screening on 42 of the high-confidence variants across multiple cell types, validating the results through a second independent method.
The payoff was concrete. For Alzheimer's disease, the team implicated three new risk genes: C4A, PVRL2, and APOC1. For progressive supranuclear palsy, they identified PLEKHM1 and KANSL1 as previously unrecognized contributors. But the discovery went deeper than simply naming new genes. The researchers found evidence that in progressive supranuclear palsy, multiple genetic variants scattered across the genome act additively—they pile on top of each other—to disrupt a specific set of transcription factors, the molecular switches that turn genes on and off. This suggested that disease risk emerges not from single broken genes but from disruption of coordinated regulatory networks operating in particular cell types.
Geschwind emphasized that the high-throughput approach, while powerful, works best when paired with careful, targeted follow-up experiments. The goal was not to replace traditional genetic investigation but to bridge a critical gap: the space between identifying disease-associated regions through population studies and understanding the actual biological mechanisms at work. "This just provides a key step between the GWAS and understanding disease mechanisms," he said. Cooper noted that combining multiple methodologies gave the team greater confidence in their findings while also illuminating how complex human genetic variation truly is.
The implications for treatment are potentially significant. If disease emerges from disruption of gene regulatory networks rather than failure of single genes, then targeting a network of genes simultaneously might prove more effective than the traditional approach of developing drugs against individual proteins. Geschwind suggested the field is entering a new era where network-based therapies become plausible. The next phase involves studying how the newly identified risk genes interact within cells and in model systems—work that will determine whether this theoretical promise translates into actual therapeutic options for patients facing these devastating diseases.
Notable Quotes
This just provides a key step between the GWAS and understanding disease mechanisms.— Dr. Dan Geschwind, UCLA
We're entering a new stage of therapies—it's beginning to be plausible to think about targeting networks.— Dr. Dan Geschwind
The Hearth Conversation Another angle on the story
Why does it matter that you can test thousands of variants at once instead of one at a time?
Because the bottleneck in genetics has always been this: you know a region of the genome is involved in disease, but that region contains hundreds of variants in tight linkage. Testing them individually would take years. Doing it in parallel collapses that timeline and lets you actually see which variants are functional.
So before this, researchers were basically throwing darts in the dark?
Not quite. They knew the general neighborhood. But imagine knowing a crime happened on a city block without knowing which house. You'd have to investigate each one separately. This lets you screen the whole block at once.
What's the significance of finding that variants work additively—piling on each other?
It changes how we think about disease risk. We've been trained to hunt for the one broken gene. But what if disease emerges when multiple small disruptions hit the same regulatory network? Then you can't fix it by targeting one gene. You need to think about the whole system.
Does that make treatment harder or easier?
Potentially easier, paradoxically. If you understand the network, you can target multiple nodes at once. It's like fixing a broken orchestra—sometimes you need to tune more than one instrument.
How confident are they in these findings?
Confident enough to publish in Science and to validate through two independent methods. But they're clear this is a beginning, not an ending. The real work—understanding how these genes actually interact in living cells—is still ahead.