Amidst growing concerns over biodiversity loss and environmental change, scientists are employing advanced computational methods to reveal the genetic and evolutionary factors that contribute to species' resilience. At the Norwegian University of Science and Technology (NTNU), researchers are at the forefront of this movement, utilizing decades of ecological data on house sparrows in northern Norway and harnessing the powerful computational capabilities of NTNU’s flagship supercomputer, IDUN. These efforts not only enhance our knowledge of evolutionary dynamics in wild populations but also provide robust quantitative tools that could inform conservation strategies for a diverse range of species.
House sparrows (Passer domesticus), though ubiquitous across much of the world, present a compelling model for studying evolution in fragmented, wild populations. Along the coast of Helgeland, archipelagos of small islands have been the site of continuous biological monitoring for over three decades. Biologists have meticulously recorded the life histories, from birth to death, of tens of thousands of individual sparrows, amassing an unparalleled dataset of genetic, morphological, and ecological measurements.
In a recent study published in Evolution, NTNU researchers applied a sophisticated statistical method known as genomic prediction (GP) to this extensive dataset, aiming to assess the accuracy of predicting genetic traits across distinct wild populations. Although widely used in agriculture and breeding programs, genomic prediction has rarely been applied within the context of wild populations due to the complexity and scale of the data.
Where observational fieldwork leaves off, supercomputing fills the gap. Kenneth Aase, a Ph.D. research fellow at NTNU’s Department of Mathematical Sciences, emphasizes that testing model assumptions and running high-dimensional simulations requires computational resources capable of handling large datasets and complex statistical models. For the most challenging computations in his analyses, Aase turns to IDUN, NTNU’s powerful HPC system, enabling large-scale simulations and hypothesis testing that would be infeasible on standard computing platforms.
Supercomputers such as IDUN provide not only raw processing power but also the ability to manage multifactorial models involving hundreds of thousands to millions of genetic markers, environmental variables, and phenotypic traits. This capability enables researchers to simulate the interaction of genetic variation and environmental pressures over time, a crucial step in understanding evolutionary trajectories in fluctuating habitats.
The insights emerging from this work extend far beyond the sparrow populations themselves. By evaluating how genomic prediction performs across separated island populations, the researchers revealed limitations and opportunities in applying such models to wild species with distinct genetic backgrounds. These findings inform not only evolutionary biology but also conservation strategies for species facing rapid environmental change.
Crucially, the computational framework developed and tested with IDUN simulations lays the groundwork for broader applications. The GPWILD project, funded by a European Research Council grant, aims to generalize these methods to other species, including Svalbard reindeer, Scottish deer, and arctic foxes, each with unique evolutionary dynamics and conservation challenges.
As climate change and habitat loss continue to exert pressure on wild populations globally, quantitative tools that couple genomic data with supercomputing-enabled modeling become indispensable. They allow scientists to evaluate adaptive potential, predict responses to environmental stressors, and identify populations at greatest risk of decline, all through simulation frameworks that capture the complex interplay of genetics and ecology.
For SC Online readers, the NTNU house sparrow initiative highlights a key insight: supercomputers now play a pivotal role beyond physical sciences and artificial intelligence, serving as powerful catalysts in evolutionary biology and conservation research. By merging decades of detailed ecological data with high-performance computing simulations and advanced statistical models, scientists are forging innovative approaches to better understand and safeguard the natural world amid rapid global change.
