Computational tracing of B cell lineages during immune responses using Dowser 3934

Abstract Description Antibodies produced by B cells drive adaptive immune responses to infections and vaccines, but also cause food allergies and many autoimmune diseases. Somatic V(D)J recombination as well as rapid somatic hypermutation and antigen-driven selection shape human antibody diversity. The complexity and dynamism of these processes make them difficult to study in humans experimentally. Single cell RNAseq can now be paired with V(D)J sequences in single B cells, enabling high throughput analysis of B cells from human blood and tissue samples. Like other evolving populations, B cell dynamics can be studied using phylogenetic lineage trees, which represent the series of somatic mutations among sampled cells. Lineage trees can link B cells from different timepoints, tissues, or subtypes. B cell trees containing multiple tissues represent trafficking between tissues, and those with multiple timepoints represent clonal evolution over time. Here, we introduce new computational methods to construct B cells lineage trees from single cell V(D)J sequences and understand patterns of B cell migration, differentiation, and evolution over time. We use this framework to decode patterns of B cell migration and differentiation during influenza vaccination and the autoimmune disease myasthenia gravis. These methods are widely applicable and are implemented in the R package dowser, available at https://dowser.readthedocs.io Funding Sources Supported by NIH R00AI159302 Topic Categories Computational and Systems Immunology (COMP)