immuneSIM: tunable multi-feature simulation of B- and T-cell receptor repertoires for immunoinformatics benchmarking

Predictability of antigen binding based on short motifs in the antibody CDRH3

Adaptive immune receptors, such as antibodies and T-cell receptors, recognize foreign threats with exquisite specificity. A major challenge in adaptive immunology is discovering the rules governing immune receptor-antigen binding in order to predict the antigen binding status of previously unseen immune receptors. Many studies assume that the antigen binding status of an immune receptor may be determined by the presence of a short motif in the complementarity determining region 3 (CDR3), disregarding other amino acids. To test this assumption, we present a method to discover short motifs which show high precision in predicting antigen binding and generalize well to unseen simulated and experimental data. Our analysis of a mutagenesis-based antibody dataset reveals 11 336 position-specific, mostly gapped motifs of 3-5 amino acids that retain high precision on independently generated experimental data. Using a subset of only 178 motifs, a simple classifier was made that on the independently generated dataset outperformed a deep learning model proposed specifically for such datasets. In conclusion, our findings support the notion that for some antibodies, antigen binding may be largely determined by a short CDR3 motif. As more experimental data emerge, our methodology could serve as a foundation for in-depth investigations into antigen binding signals.

Benchmarking and integrating human B-cell receptor genomic and antibody proteomic profiling

Immunoglobulins (Ig), which exist either as B-cell receptors (BCR) on the surface of B cells or as antibodies when secreted, play a key role in the recognition and response to antigenic threats. The capability to jointly characterize the BCR and antibody repertoire is crucial for understanding human adaptive immunity. From peripheral blood, bulk BCR sequencing (bulkBCR-seq) currently provides the highest sampling depth, single-cell BCR sequencing (scBCR-seq) allows for paired chain characterization, and antibody peptide sequencing by tandem mass spectrometry (Ab-seq) provides information on the composition of secreted antibodies in the serum. Yet, it has not been benchmarked to what extent the datasets generated by these three technologies overlap and complement each other. To address this question, we isolated peripheral blood B cells from healthy human donors and sequenced BCRs at bulk and single-cell levels, in addition to utilizing publicly available sequencing data. Integrated analysis was performed on these datasets, resolved by replicates and across individuals. Simultaneously, serum antibodies were isolated, digested with multiple proteases, and analyzed with Ab-seq. Systems immunology analysis showed high concordance in repertoire features between bulk and scBCR-seq within individuals, especially when replicates were utilized. In addition, Ab-seq identified clonotype-specific peptides using both bulk and scBCR-seq library references, demonstrating the feasibility of combining scBCR-seq and Ab-seq for reconstructing paired-chain Ig sequences from the serum antibody repertoire. Collectively, our work serves as a proof-of-principle for combining bulk sequencing, single-cell sequencing, and mass spectrometry as complementary methods towards capturing humoral immunity in its entirety.

nf-core/airrflow: An adaptive immune receptor repertoire analysis workflow employing the Immcantation framework

Adaptive Immune Receptor Repertoire sequencing (AIRR-seq) is a valuable experimental tool to study the immune state in health and following immune challenges such as infectious diseases, (auto)immune diseases, and cancer. Several tools have been developed to reconstruct B cell and T cell receptor sequences from AIRR-seq data and infer B and T cell clonal relationships. However, currently available tools offer limited parallelization across samples, scalability or portability to high-performance computing infrastructures. To address this need, we developed nf-core/airrflow, an end-to-end bulk and single-cell AIRR-seq processing workflow which integrates the Immcantation Framework following BCR and TCR sequencing data analysis best practices. The Immcantation Framework is a comprehensive toolset, which allows the processing of bulk and single-cell AIRR-seq data from raw read processing to clonal inference. nf-core/airrflow is written in Nextflow and is part of the nf-core project, which collects community contributed and curated Nextflow workflows for a wide variety of analysis tasks. We assessed the performance of nf-core/airrflow on simulated sequencing data with sequencing errors and show example results with real datasets. To demonstrate the applicability of nf-core/airrflow to the high-throughput processing of large AIRR-seq datasets, we validated and extended previously reported findings of convergent antibody responses to SARS-CoV-2 by analyzing 97 COVID-19 infected individuals and 99 healthy controls, including a mixture of bulk and single-cell sequencing datasets. Using this dataset, we extended the convergence findings to 20 additional subjects, highlighting the applicability of nf-core/airrflow to validate findings in small in-house cohorts with reanalysis of large publicly available AIRR datasets.

Best practices for machine learning in antibody discovery and development

In the past 40 years, therapeutic antibody discovery and development have advanced considerably, with machine learning (ML) offering a promising way to speed up the process by reducing costs and the number of experiments required. Recent progress in ML-guided antibody design and development (D&D) has been hindered by the diversity of data sets and evaluation methods, which makes it difficult to conduct comparisons and assess utility. Establishing standards and guidelines will be crucial for the wider adoption of ML and the advancement of the field. This perspective critically reviews current practices, highlights common pitfalls and proposes method development and evaluation guidelines for various ML-based techniques in therapeutic antibody D&D. Addressing challenges across the ML process, best practices are recommended for each stage to enhance reproducibility and progress.

For antibody sequence generative modeling, mixture models may be all you need

MOTIVATION

Antibody therapeutic candidates must exhibit not only tight binding to their target but also good developability properties, especially low risk of immunogenicity.

RESULTS

In this work, we fit a simple generative model, SAM, to sixty million human heavy and seventy million human light chains. We show that the probability of a sequence calculated by the model distinguishes human sequences from other species with the same or better accuracy on a variety of benchmark datasets containing >400 million sequences than any other model in the literature, outperforming large language models (LLMs) by large margins. SAM can humanize sequences, generate new sequences, and score sequences for humanness. It is both fast and fully interpretable. Our results highlight the importance of using simple models as baselines for protein engineering tasks. We additionally introduce a new tool for numbering antibody sequences which is orders of magnitude faster than existing tools in the literature.

AVAILABILITY AND IMPLEMENTATION

All tools developed in this study are available at https://github.com/Wang-lab-UCSD/AntPack.

Systematic evaluation of B-cell clonal family inference approaches

The reconstruction of clonal families (CFs) in B-cell receptor (BCR) repertoire analysis is a crucial step to understand the adaptive immune system and how it responds to antigens. The BCR repertoire of an individual is formed throughout life and is diverse due to several factors such as gene recombination and somatic hypermutation. The use of Adaptive Immune Receptor Repertoire sequencing (AIRR-seq) using next generation sequencing enabled the generation of full BCR repertoires that also include rare CFs. The reconstruction of CFs from AIRR-seq data is challenging and several approaches have been developed to solve this problem. Currently, most methods use the heavy chain (HC) only, as it is more variable than the light chain (LC). CF reconstruction options include the definition of appropriate sequence similarity measures, the use of shared mutations among sequences, and the possibility of reconstruction without preliminary clustering based on V- and J-gene annotation. In this study, we aimed to systematically evaluate different approaches for CF reconstruction and to determine their impact on various outcome measures such as the number of CFs derived, the size of the CFs, and the accuracy of the reconstruction. The methods were compared to each other and to a method that groups sequences based on identical junction sequences and another method that only determines subclones. We found that after accounting for data set variability, in particular sequencing depth and mutation load, the reconstruction approach has an impact on part of the outcome measures, including the number of CFs. Simulations indicate that unique junctions and subclones should not be used as substitutes for CF and that more complex methods do not outperform simpler methods. Also, we conclude that different approaches differ in their ability to correctly reconstruct CFs when not considering the LC and to identify shared CFs. The results showed the effect of different approaches on the reconstruction of CFs and highlighted the importance of choosing an appropriate method.

nf-core/airrflow: an adaptive immune receptor repertoire analysis workflow employing the Immcantation framework

Adaptive Immune Receptor Repertoire sequencing (AIRR-seq) is a valuable experimental tool to study the immune state in health and following immune challenges such as infectious diseases, (auto)immune diseases, and cancer. Several tools have been developed to reconstruct B cell and T cell receptor sequences from AIRR-seq data and infer B and T cell clonal relationships. However, currently available tools offer limited parallelization across samples, scalability or portability to high-performance computing infrastructures. To address this need, we developed nf-core/airrflow, an end-to-end bulk and single-cell AIRR-seq processing workflow which integrates the Immcantation Framework following BCR and TCR sequencing data analysis best practices. The Immcantation Framework is a comprehensive toolset, which allows the processing of bulk and single-cell AIRR-seq data from raw read processing to clonal inference. nf-core/airrflow is written in Nextflow and is part of the nf-core project, which collects community contributed and curated Nextflow workflows for a wide variety of analysis tasks. We assessed the performance of nf-core/airrflow on simulated sequencing data with sequencing errors and show example results with real datasets. To demonstrate the applicability of nf-core/airrflow to the high-throughput processing of large AIRR-seq datasets, we validated and extended previously reported findings of convergent antibody responses to SARS-CoV-2 by analyzing 97 COVID-19 infected individuals and 99 healthy controls, including a mixture of bulk and single-cell sequencing datasets. Using this dataset, we extended the convergence findings to 20 additional subjects, highlighting the applicability of nf-core/airrflow to validate findings in small in-house cohorts with reanalysis of large publicly available AIRR datasets. nf-core/airrflow is available free of charge, under the MIT license on GitHub (https://github.com/nf-core/airrflow). Detailed documentation and example results are available on the nf-core website at (https://nf-co.re/airrflow).

Large-scale data mining of four billion human antibody variable regions reveals convergence between therapeutic and natural antibodies that constrains search space for biologics drug discovery

The naïve human antibody repertoire has theoretical access to an estimated > 1015 antibodies. Identifying subsets of this prohibitively large space where therapeutically relevant antibodies may be found is useful for development of these agents. It was previously demonstrated that, despite the immense sequence space, different individuals can produce the same antibodies. It was also shown that therapeutic antibodies, which typically follow seemingly unnatural development processes, can arise independently naturally. To check for biases in how the sequence space is explored, we data mined public repositories to identify 220 bioprojects with a combined seven billion reads. Of these, we created a subset of human bioprojects that we make available as the AbNGS database (https://naturalantibody.com/ngs/). AbNGS contains 135 bioprojects with four billion productive human heavy variable region sequences and 385 million unique complementarity-determining region (CDR)-H3s. We find that 270,000 (0.07% of 385 million) unique CDR-H3s are highly public in that they occur in at least five of 135 bioprojects. Of 700 unique therapeutic CDR-H3, a total of 6% has direct matches in the small set of 270,000. This observation extends to a match between CDR-H3 and V-gene call as well. Thus, the subspace of shared ('public') CDR-H3s shows utility for serving as a starting point for therapeutic antibody design.

AIRRSHIP: simulating human B cell receptor repertoire sequences

SUMMARY

Adaptive Immune Receptor Repertoire Sequencing is a rapidly developing field that has advanced understanding of the role of the adaptive immune system in health and disease. Numerous tools have been developed to analyse the complex data produced by this technique but work to compare their accuracy and reliability has been limited. Thorough, systematic assessment of their performance is dependent on the ability to produce high quality simulated datasets with known ground truth. We have developed AIRRSHIP, a flexible and fast Python package that produces synthetic human B cell receptor sequences. AIRRSHIP uses a comprehensive set of reference data to replicate key mechanisms in the immunoglobulin recombination process, with a particular focus on junctional complexity. Repertoires generated by AIRRSHIP are highly similar to published data and all steps in the sequence generation process are recorded. These data can be used to not only determine the accuracy of repertoire analysis tools but can also, by tuning of the large number of user-controllable parameters, give insight into factors that contribute to inaccuracies in results.

AVAILABILITY AND IMPLEMENTATION

AIRRSHIP is implemented in Python. It is available via https://github.com/Cowanlab/airrship and on PyPI at https://pypi.org/project/airrship/. Documentation can be found at https://airrship.readthedocs.io/.

Immunoinformatic prediction of potential immunodominant epitopes from cagW in order to investigate protection against Helicobacter pylori infection based on experimental consequences

Helicobacter pylori is a leading cause of stomach cancer and peptic ulcers. Thus, identifying epitopes in H. pylori antigens is important for disease etiology, immunological surveillance, enhancing early detection tests, and developing optimal epitope-based vaccines. We used immunoinformatic and computational methods to create a potential CagW epitope candidate for H. pylori protection. The cagW gene of H. pylori was amplified and cloned into pcDNA3.1 (+) for injection into the muscles of healthy BALB/c mice to assess the impact of the DNA vaccine on interleukin levels. The results will be compared to a control group of mice that received PBS or cagW-pcDNA3.1 (+) vaccinations. An analysis of CagW protein antigens revealed 8 CTL and 7 HTL epitopes linked with AYY and GPGPG, which were enhanced by adding B-defensins to the N-terminus. The vaccine's immunogenicity, allergenicity, and physiochemistry were validated, and its strong activation of TLRs (1, 2, 3, 4, and 10) suggests it is antigenic. An in-silico cloning and immune response model confirmed the vaccine's expression efficiency and predicted its impact on the immune system. An immunofluorescence experiment showed stable and bioactive cagW gene expression in HDF cells after cloning the whole genome into pcDNA3.1 (+). In vivo vaccination showed that pcDNA3.1 (+)-cagW-immunized mice had stronger immune responses and a longer plasmid DNA release window than control-plasmid-immunized mice. After that, bioinformatics methods predicted, developed, and validated the three-dimensional structure. Many online services docked it with Toll-like receptors. The vaccine was refined using allergenicity, antigenicity, solubility, physicochemical properties, and molecular docking scores. Virtual-reality immune system simulations showed an impressive reaction. Codon optimization and in-silico cloning produced E. coli-expressed vaccines. This study suggests a CagW epitopes-protected H. pylori infection. These studies show that the proposed immunization may elicit particular immune responses against H. pylori, but laboratory confirmation is needed to verify its safety and immunogenicity.

Understanding repertoire sequencing data through a multiscale computational model of the germinal center

Sequencing of B-cell and T-cell immune receptor repertoires helps us to understand the adaptive immune response, although it only provides information about the clonotypes (lineages) and their frequencies and not about, for example, their affinity or antigen (Ag) specificity. To further characterize the identified clones, usually with special attention to the particularly abundant ones (dominant), additional time-consuming or expensive experiments are generally required. Here, we present an extension of a multiscale model of the germinal center (GC) that we previously developed to gain more insight in B-cell repertoires. We compare the extent that these simulated repertoires deviate from experimental repertoires established from single GCs, blood, or tissue. Our simulations show that there is a limited correlation between clonal abundance and affinity and that there is large affinity variability among same-ancestor (same-clone) subclones. Our simulations suggest that low-abundance clones and subclones, might also be of interest since they may have high affinity for the Ag. We show that the fraction of plasma cells (PCs) with high B-cell receptor (BcR) mRNA content in the GC does not significantly affect the number of dominant clones derived from single GCs by sequencing BcR mRNAs. Results from these simulations guide data interpretation and the design of follow-up experiments.

Modeling and predicting the overlap of B- and T-cell receptor repertoires in healthy and SARS-CoV-2 infected individuals

Adaptive immunity's success relies on the extraordinary diversity of protein receptors on B and T cell membranes. Despite this diversity, the existence of public receptors shared by many individuals gives hope for developing population-wide vaccines and therapeutics. Using probabilistic modeling, we show many of these public receptors are shared by chance in healthy individuals. This predictable overlap is driven not only by biases in the random generation process of receptors, as previously reported, but also by their common functional selection. However, the model underestimates sharing between repertoires of individuals infected with SARS-CoV-2, suggesting strong specific antigen-driven convergent selection. We exploit this discrepancy to identify COVID-associated receptors, which we validate against datasets of receptors with known viral specificity. We study their properties in terms of sequence features and network organization, and use them to design an accurate diagnostic tool for predicting SARS-CoV-2 status from repertoire data.

The evolutionary and functional significance of germline immunoglobulin gene variation

The recombination between immunoglobulin (IG) gene segments determines an individual's naïve antibody repertoire and, consequently, (auto)antigen recognition. Emerging evidence suggests that mammalian IG germline variation impacts humoral immune responses associated with vaccination, infection, and autoimmunity - from the molecular level of epitope specificity, up to profound changes in the architecture of antibody repertoires. These links between IG germline variants and immunophenotype raise the question on the evolutionary causes and consequences of diversity within IG loci. We discuss why the extreme diversity in IG loci remains a mystery, why resolving this is important for the design of more effective vaccines and therapeutics, and how recent evidence from multiple lines of inquiry may help us do so.

Assessing developability early in the discovery process for novel biologics

Beyond potency, a good developability profile is a key attribute of a biological drug. Selecting and screening for such attributes early in the drug development process can save resources and avoid costly late-stage failures. Here, we review some of the most important developability properties that can be assessed early on for biologics. These include the influence of the source of the biologic, its biophysical and pharmacokinetic properties, and how well it can be expressed recombinantly. We furthermore present in silico, in vitro, and in vivo methods and techniques that can be exploited at different stages of the discovery process to identify molecules with liabilities and thereby facilitate the selection of the most optimal drug leads. Finally, we reflect on the most relevant developability parameters for injectable versus orally delivered biologics and provide an outlook toward what general trends are expected to rise in the development of biologics.

simAIRR: simulation of adaptive immune repertoires with realistic receptor sequence sharing for benchmarking of immune state prediction methods

BACKGROUND

Machine learning (ML) has gained significant attention for classifying immune states in adaptive immune receptor repertoires (AIRRs) to support the advancement of immunodiagnostics and therapeutics. Simulated data are crucial for the rigorous benchmarking of AIRR-ML methods. Existing approaches to generating synthetic benchmarking datasets result in the generation of naive repertoires missing the key feature of many shared receptor sequences (selected for common antigens) found in antigen-experienced repertoires.

RESULTS

We demonstrate that a common approach to generating simulated AIRR benchmark datasets can introduce biases, which may be exploited for undesired shortcut learning by certain ML methods. To mitigate undesirable access to true signals in simulated AIRR datasets, we devised a simulation strategy (simAIRR) that constructs antigen-experienced-like repertoires with a realistic overlap of receptor sequences. simAIRR can be used for constructing AIRR-level benchmarks based on a range of assumptions (or experimental data sources) for what constitutes receptor-level immune signals. This includes the possibility of making or not making any prior assumptions regarding the similarity or commonality of immune state-associated sequences that will be used as true signals. We demonstrate the real-world realism of our proposed simulation approach by showing that basic ML strategies perform similarly on simAIRR-generated and real-world experimental AIRR datasets.

CONCLUSIONS

This study sheds light on the potential shortcut learning opportunities for ML methods that can arise with the state-of-the-art way of simulating AIRR datasets. simAIRR is available as a Python package: https://github.com/KanduriC/simAIRR.

A scalable model for simulating multi-round antibody evolution and benchmarking of clonal tree reconstruction methods

Affinity maturation (AM) of B cells through somatic hypermutations (SHMs) enables the immune system to evolve to recognize diverse pathogens. The accumulation of SHMs leads to the formation of clonal lineages of antibody-secreting b cells that have evolved from a common naïve B cell. Advances in high-throughput sequencing have enabled deep scans of B cell receptor repertoires, paving the way for reconstructing clonal trees. However, it is not clear if clonal trees, which capture microevolutionary time scales, can be reconstructed using traditional phylogenetic reconstruction methods with adequate accuracy. In fact, several clonal tree reconstruction methods have been developed to fix supposed shortcomings of phylogenetic methods. Nevertheless, no consensus has been reached regarding the relative accuracy of these methods, partially because evaluation is challenging. Benchmarking the performance of existing methods and developing better methods would both benefit from realistic models of clonal lineage evolution specifically designed for emulating B cell evolution. In this paper, we propose a model for modeling B cell clonal lineage evolution and use this model to benchmark several existing clonal tree reconstruction methods. Our model, designed to be extensible, has several features: by evolving the clonal tree and sequences simultaneously, it allows modeling selective pressure due to changes in affinity binding; it enables scalable simulations of large numbers of cells; it enables several rounds of infection by an evolving pathogen; and, it models building of memory. In addition, we also suggest a set of metrics for comparing clonal trees and measuring their properties. Our results show that while maximum likelihood phylogenetic reconstruction methods can fail to capture key features of clonal tree expansion if applied naively, a simple post-processing of their results, where short branches are contracted, leads to inferences that are better than alternative methods.

Echidna: integrated simulations of single-cell immune receptor repertoires and transcriptomes

Motivation

Single-cell sequencing now enables the recovery of full-length immune receptor repertoires [B cell receptor (BCR) and T cell receptor (TCR) repertoires], in addition to gene expression information. The feature-rich datasets produced from such experiments require extensive and diverse computational analyses, each of which can significantly influence the downstream immunological interpretations, such as clonal selection and expansion. Simulations produce validated standard datasets, where the underlying generative model can be precisely defined and furthermore perturbed to investigate specific questions of interest. Currently, there is no tool that can be used to simulate single-cell datasets incorporating immune receptor repertoires and gene expression.

Results

We developed Echidna, an R package that simulates immune receptors and transcriptomes at single-cell resolution with user-tunable parameters controlling a wide range of features such as clonal expansion, germline gene usage, somatic hypermutation, transcriptional phenotypes and spatial location. Echidna can additionally simulate time-resolved B cell evolution, producing mutational networks with complex selection histories incorporating class-switching and B cell subtype information. We demonstrated the benchmarking potential of Echidna by simulating clonal lineages and comparing the known simulated networks with those inferred from only the BCR sequences as input. Finally, we simulated immune repertoire information onto existing spatial transcriptomic experiments, thereby generating novel datasets that could be used to develop and integrate methods to profile clonal selection in a spatially resolved manner. Together, Echidna provides a framework that can incorporate experimental data to simulate single-cell immune repertoires to aid software development and bioinformatic benchmarking of clonotyping, phylogenetics, transcriptomics and machine learning strategies.

Availability and implementation

The R package and code used in this manuscript can be found at github.com/alexyermanos/echidna and also in the R package Platypus (Yermanos et al., 2021). Installation instructions and the vignette for Echidna is described in the Platypus Computational Ecosystem (https://alexyermanos.github.io/Platypus/index.html). Publicly available data and corresponding sample accession numbers can be found in Supplementary Tables S2 and S3.

Supplementary information

Supplementary data are available at Bioinformatics Advances online.

Reference-based comparison of adaptive immune receptor repertoires

B and T cell receptor (immune) repertoires can represent an individual's immune history. While current repertoire analysis methods aim to discriminate between health and disease states, they are typically based on only a limited number of parameters. Here, we introduce immuneREF: a quantitative multidimensional measure of adaptive immune repertoire (and transcriptome) similarity that allows interpretation of immune repertoire variation by relying on both repertoire features and cross-referencing of simulated and experimental datasets. To quantify immune repertoire similarity landscapes across health and disease, we applied immuneREF to >2,400 datasets from individuals with varying immune states (healthy, [autoimmune] disease, and infection). We discovered, in contrast to the current paradigm, that blood-derived immune repertoires of healthy and diseased individuals are highly similar for certain immune states, suggesting that repertoire changes to immune perturbations are less pronounced than previously thought. In conclusion, immuneREF enables the population-wide study of adaptive immune response similarity across immune states.

T-Cell Receptor Repertoire Sequencing and Its Applications: Focus on Infectious Diseases and Cancer

The immune system is a dynamic feature of each individual and a footprint of our unique internal and external exposures. Indeed, the type and level of exposure to physical and biological agents shape the development and behavior of this complex and diffuse system. Many pathological conditions depend on how our immune system responds or does not respond to a pathogen or a disease or on how the regulation of immunity is altered by the disease itself. T-cells are important players in adaptive immunity and, together with B-cells, define specificity and monitor the internal and external signals that our organism perceives through its specific receptors, TCRs and BCRs, respectively. Today, high-throughput sequencing (HTS) applied to the TCR repertoire has opened a window of opportunity to disclose T-cell repertoire development and behavior down to the clonal level. Although TCR repertoire sequencing is easily accessible today, it is important to deeply understand the available technologies for choosing the best fit for the specific experimental needs and questions. Here, we provide an updated overview of TCR repertoire sequencing strategies, providers and applications to infectious diseases and cancer to guide researchers’ choice through the multitude of available options. The possibility of extending the TCR repertoire to HLA characterization will be of pivotal importance in the near future to understand how specific HLA genes shape T-cell responses in different pathological contexts and will add a level of comprehension that was unthinkable just a few years ago.

Machine Learning Approaches to TCR Repertoire Analysis

Sparked by the development of genome sequencing technology, the quantity and quality of data handled in immunological research have been changing dramatically. Various data and database platforms are now driving the rapid progress of machine learning for immunological data analysis. Of various topics in immunology, T cell receptor repertoire analysis is one of the most important targets of machine learning for assessing the state and abnormalities of immune systems. In this paper, we review recent repertoire analysis methods based on machine learning and deep learning and discuss their prospects.

Benchmarking computational methods for B-cell receptor reconstruction from single-cell RNA-seq data

Multiple methods have recently been developed to reconstruct full-length B-cell receptors (BCRs) from single-cell RNA sequencing (scRNA-seq) data. This need emerged from the expansion of scRNA-seq techniques, the increasing interest in antibody-based drug development and the importance of BCR repertoire changes in cancer and autoimmune disease progression. However, a comprehensive assessment of performance-influencing factors such as the sequencing depth, read length or number of somatic hypermutations (SHMs) as well as guidance regarding the choice of methodology is still lacking. In this work, we evaluated the ability of six available methods to reconstruct full-length BCRs using one simulated and three experimental SMART-seq datasets. In addition, we validated that the BCRs assembled in silico recognize their intended targets when expressed as monoclonal antibodies. We observed that methods such as BALDR, BASIC and BRACER showed the best overall performance across the tested datasets and conditions, whereas only BASIC demonstrated acceptable results on very short read libraries. Furthermore, the de novo assembly-based methods BRACER and BALDR were the most accurate in reconstructing BCRs harboring different degrees of SHMs in the variable domain, while TRUST4, MiXCR and BASIC were the fastest. Finally, we propose guidelines to select the best method based on the given data characteristics.

Profiling the baseline performance and limits of machine learning models for adaptive immune receptor repertoire classification

BACKGROUND

Machine learning (ML) methodology development for the classification of immune states in adaptive immune receptor repertoires (AIRRs) has seen a recent surge of interest. However, so far, there does not exist a systematic evaluation of scenarios where classical ML methods (such as penalized logistic regression) already perform adequately for AIRR classification. This hinders investigative reorientation to those scenarios where method development of more sophisticated ML approaches may be required.

RESULTS

To identify those scenarios where a baseline ML method is able to perform well for AIRR classification, we generated a collection of synthetic AIRR benchmark data sets encompassing a wide range of data set architecture-associated and immune state-associated sequence patterns (signal) complexity. We trained ≈1,700 ML models with varying assumptions regarding immune signal on ≈1,000 data sets with a total of ≈250,000 AIRRs containing ≈46 billion TCRβ CDR3 amino acid sequences, thereby surpassing the sample sizes of current state-of-the-art AIRR-ML setups by two orders of magnitude. We found that L1-penalized logistic regression achieved high prediction accuracy even when the immune signal occurs only in 1 out of 50,000 AIR sequences.

CONCLUSIONS

We provide a reference benchmark to guide new AIRR-ML classification methodology by (i) identifying those scenarios characterized by immune signal and data set complexity, where baseline methods already achieve high prediction accuracy, and (ii) facilitating realistic expectations of the performance of AIRR-ML models given training data set properties and assumptions. Our study serves as a template for defining specialized AIRR benchmark data sets for comprehensive benchmarking of AIRR-ML methods.

Progress and challenges for the machine learning-based design of fit-for-purpose monoclonal antibodies

Although the therapeutic efficacy and commercial success of monoclonal antibodies (mAbs) are tremendous, the design and discovery of new candidates remain a time and cost-intensive endeavor. In this regard, progress in the generation of data describing antigen binding and developability, computational methodology, and artificial intelligence may pave the way for a new era of in silico on-demand immunotherapeutics design and discovery. Here, we argue that the main necessary machine learning (ML) components for an in silico mAb sequence generator are: understanding of the rules of mAb-antigen binding, capacity to modularly combine mAb design parameters, and algorithms for unconstrained parameter-driven in silico mAb sequence synthesis. We review the current progress toward the realization of these necessary components and discuss the challenges that must be overcome to allow the on-demand ML-based discovery and design of fit-for-purpose mAb therapeutic candidates.

OUP accepted manuscript

The study and manipulation of T cell receptors (TCRs) is central to multiple fields across basic and translational immunology research. Produced by V(D)J recombination, TCRs are often only recorded in the literature and data repositories as a combination of their V and J gene symbols, plus their hypervariable CDR3 amino acid sequence. However, numerous applications require full-length coding nucleotide sequences. Here we present Stitchr, a software tool developed to specifically address this limitation. Given minimal V/J/CDR3 information, Stitchr produces complete coding sequences representing a fully spliced TCR cDNA. Due to its modular design, Stitchr can be used for TCR engineering using either published germline or novel/modified variable and constant region sequences. Sequences produced by Stitchr were validated by synthesizing and transducing TCR sequences into Jurkat cells, recapitulating the expected antigen specificity of the parental TCR. Using a companion script, Thimble, we demonstrate that Stitchr can process a million TCRs in under ten minutes using a standard desktop personal computer. By systematizing the production and modification of TCR sequences, we propose that Stitchr will increase the speed, repeatability, and reproducibility of TCR research. Stitchr is available on GitHub.

In silico proof of principle of machine learning-based antibody design at unconstrained scale

Generative machine learning (ML) has been postulated to become a major driver in the computational design of antigen-specific monoclonal antibodies (mAb). However, efforts to confirm this hypothesis have been hindered by the infeasibility of testing arbitrarily large numbers of antibody sequences for their most critical design parameters: paratope, epitope, affinity, and developability. To address this challenge, we leveraged a lattice-based antibody-antigen binding simulation framework, which incorporates a wide range of physiological antibody-binding parameters. The simulation framework enables the computation of synthetic antibody-antigen 3D-structures, and it functions as an oracle for unrestricted prospective evaluation and benchmarking of antibody design parameters of ML-generated antibody sequences. We found that a deep generative model, trained exclusively on antibody sequence (one dimensional: 1D) data can be used to design conformational (three dimensional: 3D) epitope-specific antibodies, matching, or exceeding the training dataset in affinity and developability parameter value variety. Furthermore, we established a lower threshold of sequence diversity necessary for high-accuracy generative antibody ML and demonstrated that this lower threshold also holds on experimental real-world data. Finally, we show that transfer learning enables the generation of high-affinity antibody sequences from low-N training data. Our work establishes a priori feasibility and the theoretical foundation of high-throughput ML-based mAb design.

Novel Allele Detection Tool Benchmark and Application With Antibody Repertoire Sequencing Dataset

Detailed knowledge of the diverse immunoglobulin germline genes is critical for the study of humoral immunity. Hundreds of alleles have been discovered by analyzing antibody repertoire sequencing (Rep-seq or Ig-seq) data via multiple novel allele detection tools (NADTs). However, the performance of these NADTs through antibody sequences with intrinsic somatic hypermutations (SHMs) is unclear. Here, we developed a tool to simulate repertoires by integrating the full spectrum features of an antibody repertoire such as germline gene usage, junctional modification, position-specific SHM and clonal expansion based on 2152 high-quality datasets. We then systematically evaluated these NADTs using both simulated and genuine Ig-seq datasets. Finally, we applied these NADTs to 687 Ig-seq datasets and identified 43 novel allele candidates (NACs) using defined criteria. Twenty-five alleles were validated through findings of other sources. In addition to the NACs detected, our simulation tool, the results of our comparison, and the streamline of this process may benefit further humoral immunity studies via Ig-seq.

Applications of Machine and Deep Learning in Adaptive Immunity

Adaptive immunity is mediated by lymphocyte B and T cells, which respectively express a vast and diverse repertoire of B cell and T cell receptors and, in conjunction with peptide antigen presentation through major histocompatibility complexes (MHCs), can recognize and respond to pathogens and diseased cells. In recent years, advances in deep sequencing have led to a massive increase in the amount of adaptive immune receptor repertoire data; additionally, proteomics techniques have led to a wealth of data on peptide-MHC presentation. These large-scale data sets are now making it possible to train machine and deep learning models, which can be used to identify complex and high-dimensional patterns in immune repertoires. This article introduces adaptive immune repertoires and machine and deep learning related to biological sequence data and then summarizes the many applications in this field, which span from predicting the immunological status of a host to the antigen specificity of individual receptors and the engineering of immunotherapeutics.

Benchmarking of T cell receptor repertoire profiling methods reveals large systematic biases

Monitoring the T cell receptor (TCR) repertoire in health and disease can provide key insights into adaptive immune responses, but the accuracy of current TCR sequencing (TCRseq) methods is unclear. In this study, we systematically compared the results of nine commercial and academic TCRseq methods, including six rapid amplification of complementary DNA ends (RACE)-polymerase chain reaction (PCR) and three multiplex-PCR approaches, when applied to the same T cell sample. We found marked differences in accuracy and intra- and inter-method reproducibility for T cell receptor α (TRA) and T cell receptor β (TRB) TCR chains. Most methods showed a lower ability to capture TRA than TRB diversity. Low RNA input generated non-representative repertoires. Results from the 5' RACE-PCR methods were consistent among themselves but differed from the RNA-based multiplex-PCR results. Using an in silico meta-repertoire generated from 108 replicates, we found that one genomic DNA-based method and two non-unique molecular identifier (UMI) RNA-based methods were more sensitive than UMI methods in detecting rare clonotypes, despite the better clonotype quantification accuracy of the latter.

Immune Literacy: Reading, Writing, and Editing Adaptive Immunity

Advances in reading, writing, and editing DNA are providing unprecedented insights into the complexity of immunological systems. This combination of systems and synthetic biology methods is enabling the quantitative and precise understanding of molecular recognition in adaptive immunity, thus providing a framework for reprogramming immune responses for translational medicine. In this review, we will highlight state-of-the-art methods such as immune repertoire sequencing, immunoinformatics, and immunogenomic engineering and their application toward adaptive immunity. We showcase novel and interdisciplinary approaches that have the promise of transforming the design and breadth of molecular and cellular immunotherapies.