scAmpi-A versatile pipeline for single-cell RNA-seq analysis from basics to clinics.

Single-cell RNA sequencing (scRNA-seq) has emerged as a powerful technique to decipher tissue composition at the single-cell level and to inform on disease mechanisms, tumor heterogeneity, and the state of the immune microenvironment. Although multiple methods for the computational analysis of scRNA-seq data exist, their application in a clinical setting demands standardized and reproducible workflows, targeted to extract, condense, and display the clinically relevant information. To this end, we designed scAmpi (Single Cell Analysis mRNA pipeline), a workflow that facilitates scRNA-seq analysis from raw read processing to informing on sample composition, clinically relevant gene and pathway alterations, and in silico identification of personalized candidate drug treatments. We demonstrate the value of this workflow for clinical decision making in a molecular tumor board as part of a clinical study.

Learning the High-Dimensional Immunogenomic Features That Predict Public and Private Antibody Repertoires.

Recent studies have revealed that immune repertoires contain a substantial fraction of public clones, which may be defined as Ab or TCR clonal sequences shared across individuals. It has remained unclear whether public clones possess predictable sequence features that differentiate them from private clones, which are believed to be generated largely stochastically. This knowledge gap represents a lack of insight into the shaping of immune repertoire diversity. Leveraging a machine learning approach capable of capturing the high-dimensional compositional information of each clonal sequence (defined by CDR3), we detected predictive public clone and private clone-specific immunogenomic differences concentrated in CDR3's N1-D-N2 region, which allowed the prediction of public and private status with 80% accuracy in humans and mice. Our results unexpectedly demonstrate that public, as well as private, clones possess predictable high-dimensional immunogenomic features. Our support vector machine model could be trained effectively on large published datasets (3 million clonal sequences) and was sufficiently robust for public clone prediction across individuals and studies prepared with different library preparation and high-throughput sequencing protocols. In summary, we have uncovered the existence of high-dimensional immunogenomic rules that shape immune repertoire diversity in a predictable fashion. Our approach may pave the way for the construction of a comprehensive atlas of public mouse and human immune repertoires with potential applications in rational vaccine design and immunotherapeutics.

Comparison of methods for phylogenetic B-cell lineage inference using time-resolved antibody repertoire simulations (AbSim).

MOTIVATION: The evolution of antibody repertoires represents a hallmark feature of adaptive B-cell immunity. Recent advancements in high-throughput sequencing have dramatically increased the resolution to which we can measure the molecular diversity of antibody repertoires, thereby offering for the first time the possibility to capture the antigen-driven evolution of B cells. However, there does not exist a repertoire simulation framework yet that enables the comparison of commonly utilized phylogenetic methods with regard to their accuracy in inferring antibody evolution. RESULTS: Here, we developed AbSim, a time-resolved antibody repertoire simulation framework, which we exploited for testing the accuracy of methods for the phylogenetic reconstruction of B-cell lineages and antibody molecular evolution. AbSim enables the (i) simulation of intermediate stages of antibody sequence evolution and (ii) the modeling of immunologically relevant parameters such as duration of repertoire evolution, and the method and frequency of mutations. First, we validated that our repertoire simulation framework recreates replicates topological similarities observed in experimental sequencing data. Second, we leveraged Absim to show that current methods fail to a certain extent to predict the true phylogenetic tree correctly. Finally, we formulated simulation-validated guidelines for antibody evolution, which in the future will enable the development of accurate phylogenetic methods. AVAILABILITY AND IMPLEMENTATION: https://cran.r-project.org/web/packages/AbSim/index.html. CONTACT: sai.reddy@ethz.ch. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Bioinformatic and Statistical Analysis of Adaptive Immune Repertoires.

High-throughput sequencing (HTS) of immune repertoires has enabled the quantitative analysis of adaptive immune responses and offers the potential to revolutionize research in lymphocyte biology, vaccine profiling, and monoclonal antibody engineering. Advances in sequencing technology coupled to an exponential decline in sequencing costs have fueled the recent overwhelming interest in immune repertoire sequencing. This, in turn, has sparked the development of numerous methods for bioinformatic and statistics-driven interpretation and visualization of immune repertoires. Here, we review the current literature on bioinformatic and statistical analysis of immune repertoire HTS data and discuss underlying assumptions, applicability, and scope. We further highlight important directions for future research, which could propel immune repertoire HTS to becoming a standard method for measuring adaptive immune responses.

Quantitative assessment of the robustness of next-generation sequencing of antibody variable gene repertoires from immunized mice.

BACKGROUND: Next-generation sequencing (NGS) of antibody variable regions has emerged as a powerful tool in systems immunology by providing quantitative molecular information on polyclonal humoral immune responses. Reproducible and robust information on antibody repertoires is valuable for basic and applied immunology studies: thus, it is essential to establish the reliability of antibody NGS data. RESULTS: We isolated RNA from antibody-secreting cells (ASCs) from either 1 mouse or a pool of 9 immunized mice in order to simulate both normal and high diversity populations. Next, we prepared three technical replicates of antibody libraries by RT-PCR from each diversity scenario, which were sequenced using the Illumina MiSeq platform resulting in >106 250 bp paired-end reads per replicate. We then assessed the robustness of antibody repertoire data based on clonal identification defined by amino acid sequence of either full-length VDJ region or the complementarity determining region 3 (CDR3). Leveraging modeling approaches adapted from mathematical ecology, we found that in either diversity scenario both CDR3 and VDJ detection nears completeness indicating deep coverage of ASC repertoires. Additionally, we defined reliability thresholds for accurate quantification and ranking of CDR3s and VDJs. Importantly, we show that both factors-(i) replicate sequencing and (ii) sequencing depth-are crucial for robust CDR3 and VDJ detection and ranking. CONCLUSIONS: In summary, we established widely applicable experimental and computational guidelines for robust antibody NGS and analysis, which will help advance systems immunology studies related to the quantitative profiling of antibody responses following infection and vaccination.

Clinical Outcomes in Pediatric Patients With Type 1 Diabetes With Early Versus Late Diagnosis: Analysis From the DPV Registry.

OBJECTIVE: This study was conducted to evaluate the effects of early clinical diagnosis of type 1 diabetes by comparison of clinical parameters at diagnosis and during follow-up in patients with pediatric type 1 diabetes with early, intermediate, and late diagnosis. RESEARCH DESIGN AND METHODS: In a population-based analysis, data on 14,292 pediatric patients with type 1 diabetes diagnosed between 2015 and 2019 were retrieved from the Diabetes Prospective Documentation (DPV) registry in March 2023. Patients were divided into four groups: one with diabetic ketoacidosis (DKA) at diagnosis and three with early, intermediate, or late diagnosis based on age-dependent HbA1c terciles. Laboratory-measured HbA1c values and those estimated from continuous glucose monitoring were aggregated as a combined glucose indicator (CGI). Insulin-dose adjusted CGI values <9% were defined as partial remission. RESULTS: At diagnosis, patients had a median age of 9.8 years (IQR = 6.8; 13.0). Three years later, patients with early diagnosis had lower CGI than patients with late diagnosis or DKA (mean [95% CI] 7.46% [7.40; 7.53] vs. 7.81% [7.75; 7.87] or 7.74% [7.68; 7.79], respectively; each P < 0.001). More patients experienced partial remission (12.6% [11.0; 14.4] vs. 9.1% [7.7; 10.7] or 8.6% [7.3; 10.0]; each P < 0.001), and 11.7% [10.2; 13.5] of patients with intermediate diagnosis were in partial remission. CONCLUSIONS: Early clinical diagnosis of type 1 diabetes may be beneficial for metabolic control and remission after 3 years of follow-up. Patients diagnosed early may represent a distinct group with better resources or with a different disease biology and slower ß-cell destruction, which needs further evaluation.

Metabolic control during the first two years of the COVID-19 pandemic in pediatric patients with type 1 diabetes: results from the German DPV initiative.

AIM: To assess effects of the SARS-CoV2 pandemic on metabolic control in youth with type 1 diabetes (T1D) in Germany in a population-based analysis. METHODS: Data from 33,372 pediatric T1D patients from the Diabetes Prospective Follow-up (DPV) registry, with face-to-face visits or telemedicine contacts in the years 2019-2021, were available. Datasets from eight time periods between March 15, 2020, and December 31, 2021, according to SARS-CoV2 incidence waves, were compared to those from five control time periods. Parameters of metabolic control were assessed with adjustment for sex, age, diabetes duration, and repeated measurements. Laboratory-measured HbA1c values and those estimated from CGM were aggregated into a combined glucose indicator (CGI). RESULTS: There was no clinically relevant difference in metabolic control between pandemic and control time periods with adjusted CGI values ranging from 7.61% [7.60-7.63] (mean [95% confidence interval (CI)]) in the third quarter of 2019 to 7.83% [7.82-7.85] in the time period from January 1 to March 15 2020, in the other control periods, and during the pandemic, CGI values lay between these values. BMI-SDS rose during the pandemic from 0.29 [0.28-0.30] (mean [95% CI]) in the third quarter of 2019 to 0.40 [0.39-0.41] during the fourth wave. Adjusted insulin dose rose during the pandemic. Event rates for hypoglycemic coma and diabetic ketoacidosis remained unchanged. CONCLUSIONS: We found no clinically relevant change of glycemic control or incidence of acute diabetes complications during the pandemic. The observed BMI increase may represent an important health risk for youth with T1D.

Prevalence of type 1 and type 2 diabetes in children and adolescents in Germany from 2002 to 2020: A study based on electronic health record data from the DPV registry.

BACKGROUND: To provide estimates of the nationwide prevalence of type 1 diabetes (T1D) and type 2 diabetes (T2D) in individuals younger than 20 years of age in Germany from 2002 to 2020 and to identify trends. METHODS: Data were obtained from the electronic health record "Diabetes Prospective Follow-up Registry (DPV)" specific to diabetes care. Prevalence was estimated based on prevalent cases at the end of each year for the years 2002, 2008, 2014, and 2020 per 100 000 persons assuming a Poisson distribution and directly age- and/or sex-standardized to the population in 2020. Individuals younger than 20 years of age with a clinical diagnosis of T1D or 10-19-year-olds with T2D were eligible for inclusion in the study. RESULTS: The standardized T1D prevalence per 100 000 persons was 138.9 (95% CI: 137.1; 140.6) in 2002 and 245.6 (243.1; 248.0) in 2020. The standardized T2D prevalence per 100 000 persons was 3.4 (3.1; 3.8) in 2002 and 10.8 (10.1; 11.5) in 2020. The annual percent change (APC) in prevalence declined over the three periods 2002-2008/2008-2014/2014-2020 (T1D: 6.3% [3.6%; 9.0%]/3.1% [0.7%; 5.5%]/0.5% [-1.7%; 2.85], T2D: 12.3% [5.3%; 20.8%]/4.7% [-0.6%; 10.3%]/3.0% [-1.8%; 8.0%]). From 2014 to 2020, the highest APCs were observed among 15-19-year-olds (T1D: 2.5% [1.3%; 3.6%], T2D: 3.4% [-0.5%; 7.5%]). CONCLUSIONS: The increase in diabetes prevalence has slowed, but medical care should be prepared for an increase in adolescents with diabetes.

Collection and preprocessing of fine needle aspirate patient samples for single cell profiling and data analysis.

High cell viability and recovered cell concentration are typical quality control requirements for single-cell processing and quality data. This protocol describes procedures for sampling, live-cell biobanking, preprocessing for single-cell RNA sequencing, and analysis of fine-needle aspiration (FNA) samples of the skin. The minimally invasive nature of FNA collection is more accepted by patients and allows for frequent longitudinal sampling, resulting in high-quality single-cell sequencing data that capture cellular heterogeneity in clinical samples.

Oncolytic virotherapy-mediated anti-tumor response: a single-cell perspective.

Talimogene laherparepvec (T-VEC) is a genetically modified herpes simplex 1 virus (HSV-1) approved for cancer therapy. We investigate its effect on the clinical, histological, single-cell transcriptomic, and immune repertoire level using repeated fine-needle aspirates (FNAs) of injected and noninjected lesions in primary cutaneous B cell lymphoma (pCBCL). Thirteen patients received intralesional T-VEC, 11 of which demonstrate tumor response in the injected lesions. Using single-cell sequencing of the FNAs, we identify the malignant population and separate three pCBCL subtypes. Twenty-four hours after the injection, we detect HSV-1T-VEC transcripts in malignant and nonmalignant cells of the injected lesion but not of the noninjected lesion. Oncolytic virotherapy results in a rapid eradication of malignant cells. It also leads to interferon pathway activation and early influx of natural killer cells, monocytes, and dendritic cells. These events are followed by enrichment in cytotoxic T cells and a decrease of regulatory T cells in injected and noninjected lesions.

Early versus delayed insulin pump therapy in children with newly diagnosed type 1 diabetes: results from the multicentre, prospective diabetes follow-up DPV registry.

BACKGROUND: Although continuous subcutaneous insulin infusion therapy (ie, insulin pump therapy) is associated with improved metabolic control compared with multiple daily insulin injections in children with type 1 diabetes, it is unclear when it is best to start it after diagnosis. In this study, we aimed to compare the outcomes between early and delayed start of insulin pump therapy in young patients with type 1 diabetes. METHODS: We based the current study on data from the multicentre, prospective diabetes follow-up registry (ie, Diabetes-Patienten-Verlaufsdokumentation [DPV]). The DPV registry comprises 501 diabetes centres from Germany, Austria, Switzerland, and Luxembourg. We included patients diagnosed with type 1 diabetes between 2004 and 2014, who were aged between 6 months and 15 years at the time of diagnosis, who had started insulin pump therapy either within the first 6 months (ie, the early treatment group) or in the second to third year (ie, the delayed treatment group) after diabetes diagnosis, and who were treated with insulin pump therapy for at least 1 year. The outcome parameters included the glycated haemoglobin (HbA1c) values, the cardiovascular risk profile, and rates of acute complications and diabetes-associated hospital admissions (ie, hospitalisation) during the most recent documented treatment year with insulin pump therapy. Statistical models were adjusted for age at diabetes diagnosis, year of diagnosis, sex, immigrant background, use of continuous glucose monitoring, centre size, and the German Index of Socioeconomic Deprivation 2012 terciles. FINDINGS: Our study sample comprised 8332 patients from 311 diabetes centres in Germany, Austria, Switzerland, and Luxembourg. The early treatment group consisted of 4004 (48·1%) of 8332 patients, and the delayed treatment group consisted of 4328 (51·9%). The median diabetes duration during follow-up was 6·7 years (IQR 5·1-8·7 in the early group; 5·0-8·7 in the delayed group) in both groups. Patients with early initiation of insulin pump therapy compared with those with delayed initiation of insulin pump therapy had significantly lower estimated mean HbA1c values (7·9% [95% CI 7·8-7·9] and 62·6 mmol/mol [95% CI 62·1-63·2] vs 8·0% [8·0-8·1] and 64·1 mmol/mol [63·6-64·6]; p=0·0006), and lower rates of hypoglycaemic coma (incidence risk ratio 0·44 [95% CI 0·24-0·79]; p=0·0064) and hospitalisation (0·86 [95% CI 0·78-0·94]; p=0·0016). A better cardiovascular risk profile was observed in patients with early initiation of insulin pump therapy than in those with delayed initiation: an estimated mean systolic blood pressure of 117·6 mm Hg (95% CI 117·2-117·9) versus 118·5 mm Hg (118·2-118·9), p=0·0007; and HDL cholesterol of 62·8 mg/dL (95% CI 62·2-63·5) versus 60·6 mg/dL (60·0-61·2), p<0·0001; however, diastolic blood pressure; concentrations of LDL cholesterol, non-HDL cholesterol, and triglycerides; and estimated body-mass index standard deviation scores during follow-up did not differ significantly between both groups. INTERPRETATION: Our findings provide evidence for improved clinical outcomes associated with the early initiation of insulin pump therapy in children with type 1 diabetes. FUNDING: The German Center for Diabetes Research (Deutsches Zentrum für Diabetesforschung), German Robert Koch Institute, German Diabetes Association, and Diabetes Agenda 2010.

The Tumor Profiler Study: integrated, multi-omic, functional tumor profiling for clinical decision support.

The application and integration of molecular profiling technologies create novel opportunities for personalized medicine. Here, we introduce the Tumor Profiler Study, an observational trial combining a prospective diagnostic approach to assess the relevance of in-depth tumor profiling to support clinical decision-making with an exploratory approach to improve the biological understanding of the disease.

IgM Antibody Repertoire Fingerprints in Mice Are Personalized but Robust to Viral Infection Status.

Antibody repertoire sequencing provides a molecular fingerprint of current and past pathogens encountered by the immune system. Most repertoire studies in humans require measuring the B cell response in the blood, resulting in a large bias to the IgM isotype. The extent to which the circulating IgM antibody repertoire correlates to lymphoid tissue-resident B cells in the setting of viral infection remains largely uncharacterized. Therefore, we compared the IgM repertoires from both blood and bone marrow (BM) plasma cells (PCs) following acute or chronic lymphocytic choriomeningitis virus (LCMV) infection in mice. Despite previously reported serum alterations between acute and chronic infection, IgM repertoire signatures based on clonal diversity metrics, public clones, network, and phylogenetic analysis were largely unable to distinguish infection cohorts. Our findings, however, revealed mouse-specific repertoire fingerprints between the blood and PC repertoires irrespective of infection status.

Chronic Viral Infection Promotes Efficient Germinal Center B Cell Responses.

Persistent viral infections subvert key elements of adaptive immunity. To compare germinal center (GC) B cell responses in chronic and acute lymphocytic choriomeningitis virus infection, we exploit activation-induced deaminase (AID) fate-reporter mice and perform adoptive B cell transfer experiments. Chronic infection yields GC B cell responses of higher cellularity than acute infections do, higher memory B cell and antibody secreting cell output for longer periods of time, a better representation of the late B cell repertoire in serum immunoglobulin, and higher titers of protective neutralizing antibodies. GC B cells of chronically infected mice are similarly hypermutated as those emerging from acute infection. They efficiently adapt to viral escape variants and even in hypermutation-impaired AID mutant mice, chronic infection selects for GC B cells with hypermutated B cell receptors (BCRs) and neutralizing antibody formation. These findings demonstrate that, unlike for CD8+ T cells, chronic viral infection drives a functional, productive, and protective GC B cell response.

Quantitative and Qualitative Analysis of Humoral Immunity Reveals Continued and Personalized Evolution in Chronic Viral Infection.

Control of established chronic lymphocytic choriomeningitis virus (LCMV) infection requires the production of neutralizing antibodies, but it remains unknown how the ensemble of antibodies evolves during ongoing infection. Here, we analyze the evolution of antibody responses during acute or chronic LCMV infection, combining quantitative functional assays and time-resolved antibody repertoire sequencing. We establish that antibody responses initially converge in both infection types on a functional and repertoire level, but diverge later during chronic infection, showing increased clonal diversity, the appearance of mouse-specific persistent clones, and distinct phylogenetic signatures. Chronic infection is characterized by a longer-lasting germinal center reaction and a continuous differentiation of plasma cells, resulting in the emergence of higher-affinity plasma cells exhibiting increased antibody secretion rates. Taken together, our findings reveal the emergence of a personalized antibody response in chronic infection and support the concept that maintaining B cell diversity throughout chronic LCMV infection correlates with the development of infection-resolving antibodies.

Discrete populations of isotype-switched memory B lymphocytes are maintained in murine spleen and bone marrow.

At present, it is not clear how memory B lymphocytes are maintained over time, and whether only as circulating cells or also residing in particular tissues. Here we describe distinct populations of isotype-switched memory B lymphocytes (Bsm) of murine spleen and bone marrow, identified according to individual transcriptional signature and B cell receptor repertoire. A population of marginal zone-like cells is located exclusively in the spleen, while a population of quiescent Bsm is found only in the bone marrow. Three further resident populations, present in spleen and bone marrow, represent transitional and follicular B cells and B1 cells, respectively. A population representing 10-20% of spleen and bone marrow memory B cells is the only one qualifying as circulating. In the bone marrow, all cells individually dock onto VCAM1+ stromal cells and, reminiscent of resident memory T and plasma cells, are void of activation, proliferation and mobility.

A novel anti-HER2 anthracycline-based antibody-drug conjugate induces adaptive anti-tumor immunity and potentiates PD-1 blockade in breast cancer.

Increasing evidence suggests that antibody-drug conjugates (ADCs) can enhance anti-tumor immunity and improve clinical outcome. Here, we elucidate the therapeutic efficacy and immune-mediated mechanisms of a novel HER2-targeting ADC bearing a potent anthracycline derivate as payload (T-PNU) in a human HER2-expressing syngeneic breast cancer model resistant to trastuzumab and ado-trastuzumab emtansine. Mechanistically, the anthracycline component of the novel ADC induced immunogenic cell death leading to exposure and secretion of danger-associated molecular signals. RNA sequencing derived immunogenomic signatures and TCRß clonotype analysis of tumor-infiltrating lymphocytes revealed a prominent role of the adaptive immune system in the regulation of T-PNU mediated anti-cancer activity. Depletion of CD8 T cells severely reduced T-PNU efficacy, thus confirming the role of cytotoxic T cells as drivers of the T-PNU mediated anti-tumor immune response. Furthermore, T-PNU therapy promoted immunological memory formation in tumor-bearing animals protecting those from tumor rechallenge. Finally, the combination of T-PNU and checkpoint inhibition, such as α-PD1, significantly enhanced tumor eradication following the treatment. In summary, a novel PNU-armed, HER2-targeting ADC elicited long-lasting immune protection in a murine orthotopic breast cancer model resistant to other HER2-directed therapies. Our findings delineate the therapeutic potential of this novel ADC payload and support its clinical development for breast cancer patients and potentially other HER2 expressing malignancies.

Characterization of diabetes following pancreatic surgery in patients with congenital hyperinsulinism.

BACKGROUND: Congenital hyperinsulinism (CHI) is the most common cause of persistent hypoglycaemia in infancy that leads to unfavourable neurological outcome if not treated adequately. In patients with severe diffuse CHI it remains under discussion whether pancreatic surgery should be performed or intensive medical treatment with the acceptance of recurrent episodes of mild hypoglycaemia is justified. Near-total pancreatectomy is associated with high rates of insulin-dependent diabetes mellitus and exocrine pancreatic insufficiency. Little is known about the management and long-term glycaemic control of CHI patients with diabetes after pancreatic surgery. We searched the German/Austrian DPV database and compared the course of 42 CHI patients with diabetes to that of patients with type 1 diabetes mellitus (T1DM). Study groups were compared at diabetes onset and after a follow-up period of 6.1 [3.3-9.7] (median [interquartile range]) years. RESULTS: The majority of CHI patients with diabetes were treated with insulin (85.2% [70.9-99.5] at diabetes onset, and 90.5% [81.2-99.7] at follow-up). However, compared to patients with T1DM, significantly more patients in the CHI group with diabetes were treated with conventional insulin therapy (47.8% vs. 24.4%, p = 0.03 at diabetes onset, and 21.1% vs. 6.4% at follow-up, p = 0.003), and only a small number of CHI patients were treated with insulin pumps. Daily insulin dose was significantly lower in CHI patients with diabetes than in patients with T1DM, both at diabetes onset (0.3 [0.2-0.5] vs. 0.6 IE/kg/d [0.4-0.8], p = 0.003) and follow-up (0.8 [0.4-1.0] vs. 0.9 [0.7-1.0] IE/kg/d, p = 0.02), while daily carbohydrate intake was comparable in both groups. Within the first treatment year, HbA1c levels were significantly lower in CHI patients with diabetes (6.2% [5.5-7.9] vs. 7.2% [6.5-8.2], p = 0.003), but increased to a level comparable to that of T1DM patients at follow-up. Interestingly, in CHI patients, the risk of severe hypoglycaemia tends to be higher only at diabetes onset (14.8% vs. 5.8%, p = 0.1). CONCLUSIONS: In surgically treated CHI patients insulin treatment needs to be intensified in order to achieve good glycaemic control. Our data furthermore emphasize the need for improved medical treatment options for patients with diazoxide- and/or octreotide-unresponsive CHI.

Systems Analysis Reveals High Genetic and Antigen-Driven Predetermination of Antibody Repertoires throughout B Cell Development.

Antibody repertoire diversity and plasticity is crucial for broad protective immunity. Repertoires change in size and diversity across multiple B cell developmental stages and in response to antigen exposure. However, we still lack fundamental quantitative understanding of the extent to which repertoire diversity is predetermined. Therefore, we implemented a systems immunology framework for quantifying repertoire predetermination on three distinct levels: (1) B cell development (pre-B cell, naive B cell, plasma cell), (2) antigen exposure (three structurally different proteins), and (3) four antibody repertoire components (V-gene usage, clonal expansion, clonal diversity, repertoire size) extracted from antibody repertoire sequencing data (400 million reads). Across all three levels, we detected a dynamic balance of high genetic (e.g., >90% for V-gene usage and clonal expansion in naive B cells) and antigen-driven (e.g., 40% for clonal diversity in plasma cells) predetermination and stochastic variation. Our study has implications for the prediction and manipulation of humoral immunity.