GIBLE: Inferring time-resolved B cell lineage trees from spatial BCR sequencing 2257851

Abstract

Abstract Introduction The spatial dynamics of B cell affinity maturation during germinal center reactions are critical for effective immune responses. Using B cell receptor (BCR) sequences, computational models can infer the history of mutations within a B cell lineage by building phylogenetic lineage trees. These lineage trees are a powerful way to infer past B cell dynamics during infection, vaccination, and autoimmune diseases. Recently developed spatial sequencing techniques generate BCR sequences across 2D slices of tissue, but no methods exist that use this data to infer the spatial evolution of B cell lineages. Methods Here, we introduce GIBLE (Geographic Inference of B Cell Lineage Evolution), a novel Bayesian phylogenetic model that reconstructs spatial locations of unobserved, ancestral B cells, explicitly links mutation rates with physical locations, and quantifies the direction of B cell migration across tissue sections. We applied GIBLE to characterize B cell maturation in the human tonsil. Results We find that clonal lineages can seed new germinal centers, leading to parallel evolution of distinct, localized evolutionary clades. Conclusion GIBLE provides a robust and broadly applicable framework for integrating spatial and phylogenetic data to trace B cell evolution in space and time during immune responses. Funding Source National Cancer Institute grant T32CA134286 Topic Categories Computational and Systems Immunology (COMP)