Learning signatures of functional responses from immune repertoires

Abstract

The adaptive immune system is a diverse, complex system that detects, eradicates, and stores memories of the multitude of evolving pathogens encountered over a lifetime.It is composed of T cells and B cells whose receptors are generated stochastically from a large sample space via genomic recombination and selection. Upon activation, B cells can diversify their receptors further through rapid, Darwinian evolution. Though high-throughput immune repertoire sequencing technologies have enabled the unparalleled sampling of millions of adaptive immune receptor sequences, the resulting datasets are still highly sparse with regard to an individual’s immune repertoire as well as the universe of receptors that can be generated. Leveraging principled, multiscale statistical analyses robust to undersampled systems, I investigate the dynamics and differential features of adaptive immune receptors likely to be associated with functional responses to COVID-19 and post-acute sequelae of COVID-19. Moreover, I propose a method to elucidate the role of memory B cells in secondary infections by introducing a stochastic telegraph process to model the lifecycle of B cell lineages. In doing so, I provide a means to reconstruct time-resolved evolutionary histories of B cells and enable a widescale characterization of their effective rates of phenotypic switching.