In a groundbreaking development, researchers at the Institute of Science and Technology Austria (ISTA) have unveiled a new algorithm that promises to transform the analysis of biobank data. This innovation, known as genomic Vector Approximate Message Passing (gVAMP), offers unprecedented accuracy and speed in extracting and analyzing information from extensive genetic datasets. The algorithm was developed by the teams of Matthew Robinson and Marco Mondelli and is poised to advance personalized medicine significantly.
Historically, the challenge of analyzing large-scale databases such as biobanks has been daunting due to the computational costs associated with high-precision algorithms. Traditional methods required sampling entire datasets millions of times, making them impractical for real-world applications. To address this, previous approaches often sacrificed accuracy for speed. However, gVAMP breaks this mold by leveraging a mathematical framework called “approximate message passing” (AMP), which Mondelli has significantly contributed to.
Algorithmic Innovation Using Human Height
The ISTA researchers chose human height as a model complex trait to develop and test their algorithm. This decision was strategic, as human height is influenced by approximately 17 million genetic variants. The team used data from the UK Biobank, the world’s most comprehensive dataset of biological, health, and lifestyle information, to analyze these variants across hundreds of thousands of whole-genome sequences.
Al Depope, an ISTA PhD student and the study’s first author, explains, “Whereas other methods tend to analyze one snippet at a time before combining the results, gVAMP functions as a ‘joint estimation’ method. Therefore, it provides a detailed overview of the effects on a trait in the context of all variants across massive-scale genetic datasets.”
“What I find particularly important is the interpretability of our algorithm when applied in biology. In addition to allowing us to predict people’s height from their DNA more accurately than before, it also allows us to pinpoint the specific DNA regions involved,” says ISTA postdoc and co-author Jakub Bajzik.
Outperforming Existing Methods
One of the most significant achievements of gVAMP is its ability to outperform existing methods in both accuracy and processing time. To validate their approach, the researchers conducted a data simulation study, creating an artificial trait with a similar number of genetic variants as human height. This allowed them to benchmark gVAMP against other methods, demonstrating its superior performance.
“Our method achieves state-of-the-art accuracy while remaining efficient enough to perform a true joint analysis across massive-scale genetic datasets in mere days,” says Depope. “This allows us to uncover the underlying biology previously hidden by limited scale.”
From Personalized Medicine to Forensics?
The implications of gVAMP extend beyond genetic analysis. The interdisciplinary study combines expertise in information theory, mathematics, genomics, and software engineering. This diverse expertise has enabled the team to explore potential applications in personalized medicine and diagnostics. These could include predicting disease onset, severity, and symptom development, as well as extending the method to protein and epigenetic data.
Moreover, the algorithm’s potential applications in forensics are intriguing. Depope suggests that gVAMP could be used to predict a suspect’s height from DNA found at crime scenes, opening new avenues in forensic science.
“I think our algorithm might also be useful in forensics to predict a suspect’s height from the DNA found on a crime scene,” Depope notes.
The research has been published in Cell Genomics, with contributions from Al Depope, Jakub Bajzik, Marco Mondelli, and Matthew R. Robinson. The project received funding from various sources, including a Lopez-Loreta Prize, an SNSF Eccellenza Grant, and an ERC Starting Grant, with high-performance computing support from ISTA’s Scientific Service Units.
As the team continues to refine and expand the capabilities of gVAMP, the algorithm’s potential to revolutionize fields ranging from personalized medicine to forensics becomes increasingly apparent. The future of biobank data analysis looks promising, with ISTA’s innovation leading the way.
