A new high-performance computing tool called DEVIL (Differential Expression with Variational Inference Learning) enables the analysis of up to ten million cells in less than two hours, with approximately three times lower memory usage and speeds up to forty times faster than the most advanced existing methods. The study, published in Nature Communications, was developed by researchers at the University of Trieste, Area Science Park, SISSA, and Human Technopole.
The technology addresses two key challenges in modern genomics: on one hand, the computational complexity involved in analyzing enormous volumes of cellular data; on the other, the statistical difficulties arising from differences between cells taken from different patients. DEVIL therefore enables faster and more efficient analyses, reducing the risk of biased interpretations and improving the reliability of results in biomedical research.
