B Cell Clonal Lineage and Somatic Hypermutation Profiling Analysis

• Main Author: Zhou, Julian Quan
• Format: Dissertation
• Language: English
• Published: ProQuest Dissertations & Theses 01-01-2020
• Subjects: Bioinformatics; Genomics; Immunology; Influenza; Mutation; Polymerase chain reaction; Viruses
• ISBN: 9798538111473

The growing volume of B cell receptor (BCR) data from high-throughput, bulk and single-cell, adaptive immune receptor repertoire sequencing (AIRR-seq) empowers us to continue to improve our understanding of B cell-mediated immune responses. Upon encountering its cognate antigen, a B cell may become activated and seed a germinal center (GC) reaction. Within the GC, B cells rapidly proliferate and undergo diversification through somatic hypermutation (SHM). AID-initiated SHM involves error-prone repair pathways and is known for exhibiting intrinsic biases as to where mutations are introduced. As SHM is the driving force of affinity maturation, it is important to understand and capture its intrinsic biases. Current computational models often assume that the probability of the same DNA motif being targeted by SHM is the same regardless of its position along an immunoglobulin sequence. In this dissertation, we tested the validity of that assumption and found position-dependent differential SHM targeting for more than three quarters of the motifs analyzed. These included classical SHM hot spots, suggesting that the “hotness” of a classical hot spot may also depend on its position. Our findings provide insights for future computational efforts towards modeling SHM. As B cells proliferate and undergo SHM, they form clones consisting of cells which originated from the same V(D)J recombinant events, yet whose BCRs differ at the nucleotide level. Inferring the clonal relationships between B cells based on their BCRs is a fundamental problem. Historically, clonal inference methods were developed for heavy chain BCRs from bulk AIRR-seq, during which heavy:light chain pairing is lost. It has been reasoned that the heavy chain junctional diversity is sufficiently high such that, even without light chain information, the likelihood of clonally unrelated cells being clustered together will be negligibly small. Using single-cell heavy:light paired BCR data, we rigorously tested this reasoning and concluded that heavy chain-based clonal clustering is sufficient to identify most clonal relationships with confidence. This finding is especially important for studies in which heavy:light unpaired BCR data from bulk AIRR-seq need to be analyzed together with heavy:light paired BCR data from single-cell AIRR-seq. Computational tools for clonal lineage and SHM profiling analysis can help to characterize B cell responses. Combining computational analysis with experimental techniques, we and collaborators investigated whether vaccination can induce a GC reaction in human B cell response to influenza. We showed that a GC reaction can be induced; and that vaccine-induced GC B cell clones overlapping with the memory recall responses displayed high levels of SHM and were enriched for broadly cross-reactive antibodies, whereas clones unique to the GC displayed significantly lower levels of SHM and were enriched for strain-specific antibodies. Our findings suggest that GC response can play a critical role in broadening the spectrum of influenza vaccine-induced protective antibodies.