The adaptive immune system depends on an extraordinary diversity of antigen receptors. T cells and B cells generate their receptors through a process called V(D)J recombination, in which variable, diversity, and joining gene segments are randomly stitched together, with additional nucleotide insertions and deletions at the junctions. This combinatorial and junctional diversity, together with the pairing of heavy and light chains in B cells and alpha and beta chains in T cells, allows the immune system to recognize an almost limitless range of pathogens. The composition of these repertoires reflects immune development, tissue microenvironment, antigen exposure history, and disease status, making their systematic characterization essential for understanding immune surveillance and tissue-specific immune regulation.

Previous studies of the porcine immune repertoire relied largely on bulk sequencing, which averages signals across mixed cell populations and offers limited resolution for dissecting cellular heterogeneity and lineage relationships. Although high-throughput repertoire sequencing had revealed preferential V(D)J gene usage and clonotype expansion in pigs, those studies generally lacked simultaneous transcriptomic information, meaning receptor features could not be directly linked to immune cell identity, functional state, or tissue location. The new work closes that gap by integrating single-cell RNA sequencing with paired V(D)J repertoire sequencing at single-cell resolution.

To build the atlas, the team, led by researchers from Sun Yat-Sen University, the Harbin Veterinary Research Institute, and Beijing BestopCell, sampled three clinically healthy, 14-day-old male DLY crossbred piglets weighing approximately 4 kilograms. The animals were screened by RT-qPCR for major swine pathogens, including porcine epidemic diarrhea virus, transmissible gastroenteritis virus, porcine deltacoronavirus, rotavirus, porcine reproductive and respiratory syndrome virus, and African swine fever virus, and all tests were negative. None of the piglets had been vaccinated, ensuring that the recorded immune landscape reflected a true physiological baseline rather than an antigen-driven response.

The researchers collected eleven immune-associated tissues spanning lymphoid, mucosal, and circulatory compartments: peripheral blood, spleen, thymus, tonsil, lung, bone marrow, intestinal lymph node, bronchial lymph node, Peyer’s patch, and the mucous membranes of the ileum and jejunum. After careful tissue dissociation, debris removal, and quality control filtering that eliminated 22,141 low-quality cells, a total of 71,731 cells were retained for analysis. Sequencing was performed on an MGI DNBSEQ-T7 platform, and the data were processed with the Seurat and scRepertoire frameworks against the Sscrofa11.1 reference genome.

The cellular inventory revealed the expected cast of immune players: T cells, B cells, natural killer cells, plasma cells, monocytes, and neutrophils, alongside structural endothelial, epithelial, and fibroblast populations. Cell-type identities were anchored to canonical markers, with B cells expressing CD19, MS4A1, and CD79A; T cells expressing CD3D, CD3E, and CD3G; and plasma cells marked by MZB1 and CD38. Reclustering resolved five major B-cell subpopulations, including naive, memory, germinal center, plasma, and cycling plasma cells, and nine T-cell subpopulations ranging from naive and memory CD4 and CD8 subsets to regulatory T cells, cycling T cells, and double-positive CD4-CD8 cells. The distribution of these subsets varied sharply by tissue: lymphoid organs were dominated by adaptive immune cells, mucosal tissues carried more epithelial and innate components, and double-positive T cells concentrated in thymic compartments, consistent with thymocyte development.

On the repertoire side, the team identified 3,275 BCR clones and 13,431 TCR clones. Both receptor classes showed strongly biased, non-random V(D)J gene usage. The five most frequent BCR V genes, including IGHV1S201 and IGHV1-1501, together accounted for 39.41 percent of total V-gene usage, while the IGHJ501 segment alone represented 44.95 percent of J-gene usage. In the TCR repertoire, segments such as TRAV15 and TRBV20-201 were preferentially enriched, alongside dominant TRBJ genes. Complementarity-determining region 3 (CDR3) lengths, the hypervariable loops that contact antigen, fell into narrow, conserved intervals: 10 to 12 and 29 to 31 amino acids for BCRs, and 12 to 14 and 26 to 28 for TCRs, with these patterns remaining stable across tissues, reflecting structural constraints on receptor generation.

Cross-tissue analyses revealed a striking asymmetry between humoral and cellular immunity. BCR clonotypes were shared far more extensively between tissues, particularly among Peyer’s patch, intestinal lymph node, ileum mucosa, and peripheral blood, whereas TCR clones showed higher tissue specificity, with the strongest sharing between peripheral blood and spleen. Because the samples came from unchallenged animals, the authors interpret this multi-organ sharing not as antigen-driven expansion but as a manifestation of intrinsic V(D)J recombination bias, in which certain gene combinations and CDR3 lengths are preferentially generated, producing ‘public-like’ clonotypes that coexist systemically. The team also notes that B cells routinely recirculate through blood and lymph, while T cells tend toward tissue residency, though their homing-receptor analysis did not show a statistically significant difference between the two lineages, leaving the mechanistic basis open.

Mapping receptor sequences back onto the single-cell transcriptomic map through matched cell barcodes linked clonal dynamics to cell state. BCRs concentrated in plasma cell and cycling plasma cell populations, and the plasma cell subset showed enriched use of IGHA and IGHG heavy chains with elevated IGHG subclass diversity, hallmarks of antibody maturation and class-switch recombination. Immunoglobulin M accounted for more than half of heavy-chain usage in each B-cell cluster, while kappa and lambda light-chain preferences differed between subsets. TCRs were concentrated in naive CD4, naive CD8, regulatory, and memory CD4 T cells, and most TCR clonotypes existed as singletons, underscoring the enormous baseline diversity maintained even without acute immune challenge.

A key technical contribution came from the pig-specific V(D)J primers designed for the study. Compared with previously published porcine repertoire datasets, the strategy detected substantially more heavy-chain V, light-chain V, and light-chain J immunoglobulin alleles, offering improved sensitivity and broader coverage of porcine antibody diversity. The authors caution that the pooled-sample design cannot assess individual variability, that cell annotations lack independent orthogonal validation, and that antigen specificity was not directly tested. Even so, the atlas stands as a foundational reference for the healthy porcine adaptive immune system, with immediate relevance for vaccine design, infectious disease research, breeding for disease resistance, and the growing use of pigs as translational models in biomedical science.

Subject of Research: Single-cell profiling of T cell and B cell receptor V(D)J immune repertoires across tissues in healthy pigs

Article Title: A single-cell profile of the porcine immune repertoire: concurrent profiling of T cell receptor and B cell receptor V(D)J sequences in health

Article References: Zhang, H., Zhang, Q., Zhang, H., Wang, P., Geng, R., Pan, R., Li, C., Shi, F., Liu, J., Wu, Y., & Cao, Y. (2026). A single-cell profile of the porcine immune repertoire: concurrent profiling of T cell receptor and B cell receptor V(D)J sequences in health. Advanced Biotechnology, 4(4), Article 38. https://doi.org/10.1007/s44307-026-00139-2

Image Credits: AI Generated

DOI: 10.1007/s44307-026-00139-2

Keywords: single-cell RNA sequencing, V(D)J recombination, immune repertoire, T cell receptor, B cell receptor, porcine immunology, CDR3, clonotype sharing, plasma cells, DLY piglets, immunoglobulin isotypes, veterinary biotechnology

To understand why this matters, it helps to consider how T cells work. Each T cell in the body carries a receptor, known as a T-cell receptor or TCR, that recognizes fragments of foreign or abnormal proteins displayed on the surface of other cells. These fragments, called epitopes, are presented by human leukocyte antigen molecules, or HLAs, forming a peptide-HLA complex that the TCR must read with exquisite specificity. The portion of the receptor that does most of this reading is a loop called the CDR3 region on the beta chain of the receptor. Because the DNA sequences encoding CDR3 loops are assembled randomly during T-cell development, the human body can theoretically produce an astronomically large repertoire of receptors, allowing it to recognize nearly any pathogen or tumor marker it might encounter.

The problem for immunotherapy developers is that finding the right receptor for a given target is extraordinarily difficult. When a patient’s tumor displays a particular peptide on its HLA molecules, researchers want to identify or engineer TCRs that will bind that peptide-HLA complex specifically, so that engineered T cells can home in on the cancer. Today, that process depends heavily on laborious and expensive experimental screening: collecting T cells that happen to respond to the target, sequencing their receptors, cloning candidates, and testing them one by one. The sheer scale of possible CDR3 sequences means that even large screens sample only a vanishingly small corner of the search space, and promising receptors may be missed entirely. This bottleneck is precisely what the ERTransformer was designed to relieve.

The architecture of the new model rests on two pre-trained transformer networks, the same class of deep learning architecture that underlies modern large language models. The first component, EpitopeBERT, was trained on approximately 1.9 million epitope sequences, allowing it to internalize the statistical grammar of the peptide fragments that the immune system encounters. The second, ReceptorBERT, was trained on roughly 33.1 million TCR sequences, giving it a deep understanding of the structural and compositional constraints that govern which receptor loops are biologically plausible. By combining these two pre-trained encoders, ERTransformer can be conditioned on a specific epitope and then generate candidate beta-chain CDR3 sequences that are plausible as receptors while being directed toward that particular target, much as a language model conditioned on a topic generates relevant text.

Conditioning is the crucial trick. A generative model trained only on receptor sequences could produce endless plausible-looking CDR3 loops, but with no guarantee that any of them would recognize the intended target. By feeding epitope information into the generation process, the researchers steered the model so that its outputs were biased toward receptors relevant to a defined peptide-HLA context. This epitope-conditioned generation transforms the model from a random sequence generator into a targeted design tool, capable of expanding and prioritizing candidate receptors for downstream experimental screening in defined peptide-HLA and TCR-chain contexts, as the authors describe in their conclusions.

To demonstrate the capability, the team put the model through a concrete test. They selected five epitopes for which natural TCRs are already known, and asked ERTransformer to generate 1,000 candidate beta-chain CDR3 sequences for each one. The resulting pool of 5,000 artificial sequences was then analyzed for how well it satisfied the statistical expectations of real receptor biology. The generated candidates showed low sequence similarity to natural TCR beta chains, meaning the model was not simply memorizing and copying receptors from its training data, while at the same time retaining plausible CDR3 lengths, realistic amino-acid composition, and the conservative substitution patterns characteristic of genuine receptor loops. In other words, the artificial sequences looked novel but not alien, a balance that is notoriously hard to strike in generative protein design.

Novelty alone, however, proves nothing in immunology. A receptor sequence that looks plausible on paper may fail to fold, fail to bind, or bind the wrong things entirely. That is why the most striking part of the study is its wet-laboratory validation. The researchers selected a subset of the artificial TCRs and tested them experimentally using flow cytometry in defined TCR and peptide-HLA contexts, measuring the degree of T-cell activation each engineered receptor produced. The results exceeded what many in the field might have expected: the level of T-cell activation induced by the selected artificial receptors was either comparable to or even surpassed that of the natural receptors specific for the same epitopes. Synthetic sequences, dreamed up by a transformer model and never evolved in any organism’s thymus, could genuinely do the job.

The implications of that finding ripple across several areas of medicine. TCR-based immunotherapies, in which a patient’s T cells are engineered to express a receptor targeting a tumor-associated peptide, represent one of the most promising frontiers in cancer treatment. But their development has been constrained by the difficulty of sourcing high-affinity, specific receptors, particularly for targets where natural T cells respond weakly or not at all. A generative model that can propose thousands of candidate receptors for a defined target, each statistically vetted for plausibility, could dramatically expand the starting pool for screening and increase the odds of finding a receptor with the right combination of potency and specificity. The authors position ERTransformer explicitly as a tool to expand and prioritize candidates for downstream experimental screening, not as a replacement for experimentation, but as a way to make that experimentation far more efficient.

The study also illustrates a broader trend in computational biology: the power of pre-training on massive unlabeled datasets before fine-tuning on specific tasks. Just as language models learn general grammar from billions of sentences before being adapted to translation or summarization, EpitopeBERT and ReceptorBERT learned the general statistics of epitopes and receptors from millions of sequences before being combined for epitope-conditioned generation. The supplementary analyses accompanying the paper, including leakage-controlled evaluations, length distribution comparisons, and diversity and novelty analyses, suggest the team took seriously the methodological pitfalls that can make generative models appear better than they are, such as accidentally training on sequences that also appear in the test set.

Published as open access on 2 September 2026 in Genome Biology, with the work supported in part by the National Natural Science Foundation of China, the study arrives at a moment when generative artificial intelligence is reshaping protein science at large. From protein structure prediction to de novo enzyme design, machine learning models are increasingly proposing molecules that experimenters then confirm in the lab. The ERTransformer extends that paradigm to the adaptive immune system, one of the most complex and personalized molecular landscapes in biology. If the approach generalizes beyond the five epitopes tested here, the painstaking hunt for therapeutic T-cell receptors could become a collaboration between vast pre-trained models and targeted experiments, with algorithms proposing and laboratories verifying. For patients waiting on the next generation of engineered T-cell therapies, that collaboration could mean the difference between candidates found by chance and candidates found by design.

Subject of Research: Epitope-conditioned generative AI for designing T-cell receptor beta-chain CDR3 sequences for immunotherapy

Article Title: Epitope-conditioned generation of T-cell receptor β-chain CDR3 candidates using a pre-trained transformer model

Article References: Yang, J., He, B., Guan, L., Chen, S., Zhao, Y., Jiang, F., Wang, Z., Guo, Y., Xu, Z., Yuan, B., Song, J., Zhang, Q., Li, T., & Yao, J. (2026). Epitope-conditioned generation of T-cell receptor β-chain CDR3 candidates using a pre-trained transformer model. Genome Biology. https://doi.org/10.1186/s13059-026-04266-w

Image Credits: AI Generated

DOI: 10.1186/s13059-026-04266-w

Keywords: T-cell receptor, CDR3, epitope, transformer model, generative AI, immunotherapy, peptide-HLA, protein design, flow cytometry, pre-trained language model, Genome Biology, T-cell activation