Sparse Linear Algebra for Biobank-Scale Population Genetics

Speaker: Giulia Guidi, USI

Abstract: Biobank initiatives include half a million phased genomes, outpacing the tools available to analyze them. Conventional genotype formats treat variants independently, ignoring the shared ancestry structure underlying population genetic variation. Prior work introduced the Genotype Representation Graph, a compact graph encoding that captures shared ancestry directly but relies on traversal algorithms that map poorly onto modern accelerators. Here, we show that this computation can be expressed exactly as sparse linear algebra: under a topological ordering, the graph’s structure becomes strictly triangular, so the core genotype matrix–vector product reduces to a sparse triangular solve. Further, we decompose this solve into a pipelined sequence of blocked sparse matrix-vector multiplications, exposing fine-grained parallelism and enabling mature, hardware-optimized primitives in place of hand-tuned kernels. The result is substantial, portable speedups across accelerator hardware, from cloud deployments to supercomputers.

Biography: Giulia Guidi is an Assistant Professor of Computer Science at Cornell University, with affiliate appointments in Electrical and Computer Engineering, Computational Biology, and Applied Mathematics. Guidi holds a PhD in Computer Science from UC Berkeley and an MSc and BSc in Biomedical Engineering from Politecnico di Milano. Guidi’s research focuses on high-performance irregular and sparse linear algebra algorithms for large-scale computational science; her work has been recognized as a 2022 Gordon Bell Prize finalist and with the 2024 SIAM SIAG/Supercomputing Early Career Prize and the 2023 ISSNAF Young Investigator Award. Guidi received an NSF CAREER Award in 2026.

Host: Prof. Olaf Schenk