Imagine trying to read every word of a million-page book using a slow, flickering flashlight. That’s how hard it’s been for scientists to study how genes are turned on across thousands of people’s tissues. The data is massive. The tools are clunky. The wait can take weeks or months.
But now, a new software called GROMTools is like switching to a high-beam headlamp powered by a tiny battery. It can do the same job in hours instead of weeks. And it runs on regular computer systems, not just supercomputers.
Genes don’t act the same in every person or every tissue. Some genes are more active in the liver, others in the brain. How much a gene is turned on—called gene expression—can affect disease risk. Scientists use a method called GReX imputation to predict this activity from DNA alone. It helps link genetic changes to conditions like diabetes, heart disease, and mental illness.
But until now, doing this for hundreds of thousands of people was too slow and expensive.
Old tools were built for a smaller world
They needed huge amounts of memory and time. Running them in cloud labs or shared systems often meant crashing servers or paying sky-high bills.
But here's the twist: GROMTools was built for today’s era of mega-biobanks. These are giant health databases with genetic and medical records from hundreds of thousands of people. The UK Biobank, for example, holds data from half a million volunteers.
A switch that handles data smarter Think of gene prediction like assembling a car in a factory. Old tools brought in every single part—even the ones not needed—and tried to build the whole thing at once. GROMTools only grabs the parts it needs, when it needs them. It uses “sparse prediction weights,” which means it skips over genetic noise and focuses on the key switches that control gene activity.
It also reads genetic data in a stream—like a video buffering as you watch—instead of loading everything at once. The result? Smaller memory use and much faster processing.
And the output is compact. Instead of saving giant files, it stores only what’s essential, like compressing a photo without losing quality.
Predictions match old tools but run 100 times faster
In tests using UK Biobank data, GROMTools analyzed chromosome 1 across up to 450,000 people. It looked at nearly 12,000 gene-tissue pairs from 32 different cell types. The predictions were nearly identical to those from standard tools like PrediXcan and PLINK2. The match was so close that the correlation was above 0.999—almost perfect.
But the speed difference was massive. GROMTools used about 100 times less computing time and 33 times less peak memory.
That’s like mailing a letter across the country instead of waiting for a horse-drawn wagon.
This doesn't mean this treatment is available yet.
Who benefits? For now, it’s researchers. Teams studying how genes affect Alzheimer’s, cancer, or autoimmune diseases can now run these analyses routinely. What once took a month can now take a day. Labs with limited budgets can join the hunt.
But the mice didn't tell the whole story This isn’t a medical treatment. It’s a behind-the-scenes tool. Patients won’t take a pill or get a new scan because of GROMTools tomorrow. But every major medical advance starts with research. Faster tools mean faster insights.
Experts say this kind of engineering leap is just as important as new biological discoveries. Without tools that can keep up with data growth, science slows down.
What this means for you If you’ve donated DNA to a research project, this kind of tool helps scientists get more from your sample. It may lead to better risk scores for diseases or help explain why some people respond to drugs and others don’t.
But there's a catch. GROMTools is open-source and free, but it still requires technical skill to use. It’s not a push-button app. And it’s only been tested on certain genetic models and data formats.
Also, while it’s fast, it’s still limited to what current gene expression models can predict. Some tissues and rare cell types aren’t well covered yet.
The real-world impact depends on how widely it’s adopted and how well it works across different biobanks beyond the UK.
What happens next The team plans to expand GROMTools to cover all chromosomes and integrate single-cell data more deeply. Other labs are already starting to use it. As more researchers adopt it, we could see a wave of new gene-disease links published in the coming years.
Science moves fast when the right tools are in place. GROMTools may not be a household name. But it could help bring the next generation of personalized medicine one step closer.