Despite advances in information extraction driven by deep learning and large language models, performance gaps remain in highly specialized biomedical fields, where domainspecific complexity poses challenges for generalist models. In this work, we focus on the domain of autoimmunity, where the main entities of interest are autoimmune diseases, autoantibodies (i.e., molecules that may mark or cause these diseases), their molecular targets, their location in the body, and their associated clinical signs. Herein, we present AAbAAC (AutoAntibodies and Autoimmunity Annotated Corpus), a corpus of 115 abstracts selected from PubMed, where we manually annotated entities and their relationships. First, AAbAAC was used to evaluate several methods on the task of named entity recognition (NER), and secondly, to fine-tune NER models. Our study demonstrates the utility of AAbAAC for information extraction in the domain of autoimmunity, showing expected improvement in NER performance after finetuning. This illustrates the value of small-scale annotation efforts for specialized domains and contributes to the computational study of autoimmunity. The AAbAAC corpus is available at https://github.com/
We previously demonstrated that the abrogation of the ICOS pathway prevents type 1 diabetes development in the Non Obese Diabetic (NOD) mouse, but results in a CD4+ T-cell dependent autoimmune neuromyopathy in aged mice. Pancreatic islet infiltrates in conventional NOD mice and neuromuscular infiltrates in Icosl-/- NOD mice have in common that they exhibit a strong enrichment in CD4+TIGIT+ T-cells, whilst TIGIT expression in the peripheral CD4+ T-cells is limited to the CD4+FoxP3+ T-cell population.When deleting Tigit on the NOD background, diabetes incidence was found increased. Peripheral CD4+CD226+ effector T-cells exhibited an increased frequency of IL-17 producing CD4+CD226+RORgt+ T-cells versus a decreased frequency of IFN$γ$-producing CD4+CD226+Tbet+ T-cells. ICOS is expressed in both CD4+FoxP3+ and CD4+CD226+ splenic T-cell subsets. Icosl deletion leads to a decrease of CD4+FoxP3+ cells, with decrease of PD1 but increase of ICOS and CCRX3. Also in the Icosl-/- model, CD4+CD226+ T-cells are decreased by Tigit deletion, and showed an increase of CD4+CD226+RORgt+ T-cells and a decrease of CD4+CD226+Tbet+ T-cells.However, deletion of Tigit in aged Icosl-/- NOD mice population d
Some prokaryotes possess CRISPR-Cas systems that provide adaptive immunity to viruses guided by DNA segments called spacers acquired from invading phage. However, the patchy incidence and limited memory breadth of CRISPR-Cas systems suggest that their fitness benefits are offset by costs. Here, we propose that cross-reactive CRISPR targeting can lead to heterologous autoimmunity, whereby foreign spacers guide self-targeting in a spacer-length dependent fashion. Balancing antiviral defense against autoimmunity predicts a scaling relation between spacer length and CRISPR repertoire size. We find evidence for this scaling through comparative analysis of sequenced prokaryotic genomes, and show that this association also holds at the level of CRISPR types. In contrast, the scaling is absent in strains with nonfunctional CRISPR loci. Finally, we demonstrate that stochastic spacer loss can explain variations around the scaling relation, even between strains of the same species. Our results suggest that heterologous autoimmunity is a selective factor shaping the evolution of CRISPR-Cas systems.
In spite of their many facets, the phenomena of autoimmunity and immunodeficiency seem to be related to each other through the subtle links connecting retroviral mutation and action to immune response and adaptation. In a previous work, we introduced a network model of how a set of interrelated genotypes (called a quasispecies, in the stationary state) and a set of interrelated idiotypes (an idiotypic network) interact. That model, which does not cover the case of a retroviral quasispecies, was instrumental for the study of quasispecies survival when confronting the immune system and led to the conclusion that, unlike what happens when a quasispecies is left to evolve by itself, letting genotypes mutate too infrequently leads to the destruction of the quasispecies. Here we extend that genotype-idiotype interaction model by the addition of a further parameter ($ν$) to account for the action of retroviruses (i.e., the destruction of idiotypes by genotypes). We give simulation results within a suitable parameter niche, highlighting the issues of quasispecies survival and of the onset of autoimmunity through the appearance of the so-called pathogenic idiotypes. Our main findings refer
It has been known for some time that human autoimmune diseases can be triggered by viral infections. Several possible mechanisms of interactions between a virus and immune system have been analysed, with a prevailing opinion being that the onset of autoimmunity can in many cases be attributed to "molecular mimicry", where linear peptide epitopes, processed from viral proteins, mimic normal host self proteins, thus leading to a cross-reaction of immune response against virus with host cells. In this paper we present a mathematical model for the dynamics of an immune response to a viral infection and autoimmunity, which takes into account T cells with different activation thresholds. We show how the infection can be cleared by the immune system, as well as how it can lead to a chronic infection or recurrent infection with relapses and remissions. Numerical simulations of the model are performed to illustrate various dynamical regimes, as well as to analyse the potential impact of treatment of autoimmune disease in the chronic and recurrent states. The results provide good qualitative agreement with available data on immune responses to viral infections and progression of autoimmune d
This paper investigates the dynamics of immune response and autoimmunity with particular emphasis on the role of regulatory T cells (Tregs), T cells with different activation thresholds, and cytokines in mediating T cell activity. Analysis of the steady states yields parameter regions corresponding to regimes of normal clearance of viral infection, chronic infection, or autoimmune behaviour, and the boundaries of stability and bifurcations of relevant steady states are found in terms of system parameters. Numerical simulations are performed to illustrate different dynamical scenarios, and to identify basins of attraction of different steady states and periodic solutions, highlighting the important role played by the initial conditions in determining the outcome of immune interactions.
We consider self-tolerance and its failure -autoimmunity- in a minimal mathematical model of the idiotypic network. A node in the network represents a clone of B-lymphocytes and its antibodies of the same idiotype which is encoded by a bitstring. The links between nodes represent possible interactions between clones of almost complementary idiotype. A clone survives only if the number of populated neighbored nodes is neither too small nor too large. The dynamics is driven by the influx of lymphocytes with randomly generated idiotype from the bone marrow. Previous work has revealed that the network evolves towards a highly organized modular architecture, characterized by groups of nodes which share statistical properties. The structural properties of the architecture can be described analytically, the statistical properties determined from simulations are confirmed by a modular mean-field theory. To model the presence of self we permanently occupy one or several nodes. These nodes influence their linked neighbors, the autoreactive clones, but are themselves not affected by idiotypic interactions. The architecture is very similar to the case without self, but organized such that the
A major contribution to the onset and development of autoimmune disease is known to come from infections. An important practical problem is identifying the precise mechanism by which the breakdown of immune tolerance as a result of immune response to infection leads to autoimmunity. In this paper, we develop a mathematical model of immune response to a viral infection, which includes T cells with different activation thresholds, regulatory T cells (Tregs), and~a cytokine mediating immune dynamics. Particular emphasis is made on the role of time delays associated with the processes of infection and mounting the immune response. Stability analysis of various steady states of the model allows us to identify parameter regions associated with different types of immune behaviour, such as, normal clearance of infection, chronic infection, and autoimmune dynamics. Numerical simulations are used to illustrate different dynamical regimes, and to identify basins of attraction of different dynamical states. An important result of the analysis is that not only the parameters of the system, but also the initial level of infection and the initial state of the immune system determine the progress
Although Vitamin D is believed to be involved in a large number of immune responses our understanding of these processes at cellular level remained at infancy. We develop and solve a coarse grained kinetic network model to quantify the effects of variation of vitamin D on human immunity. The system of equations accounts for known inter-relation between active and inactive vitamin D, antigen presenting cells, effector T cells, regulatory T cells and pathogen. Both time dependent and steady state solutions are obtained. The time dependent solution of the system of equations reveals that the immune response is rather strongly regulated in presence of vitamin D. We found quantitatively that lower than optimum levels of concentration of active vitamin D correspond to weak regulation where, once a pathogen/antigen enters the body, the nature of the immune response would be less regulatory and hence more aggressive, or inflammatory. The steady state solution of our model shows that vitamin D enhances the tolerance level of immune system, thereby increasing resistance to autoimmune diseases. Our model and accompanied numerical analyses reveal another important aspect of immunity: While ext
T cells are a critical component of the adaptive immune system, playing a role in infectious disease, autoimmunity, and cancer. T cell function is mediated by the T cell receptor (TCR) protein, a highly diverse receptor targeting specific peptides presented by the major histocompatibility complex (pMHCs). Predicting the specificity of TCRs for their cognate pMHCs is central to understanding adaptive immunity and enabling personalized therapies. However, accurate prediction of this protein-protein interaction remains challenging due to the extreme diversity of both TCRs and pMHCs. Here, we present ImmSET (Immune Synapse Encoding Transformer), a novel sequence-based architecture designed to model interactions among sets of variable-length biological sequences. We train this model across a range of dataset sizes and compositions and study the resulting models' generalization to pMHC targets. We describe a failure mode in prior sequence-based approaches that inflates previously reported performance on this task and show that ImmSET remains robust under stricter evaluation. In systematically testing the scaling behavior of ImmSET with training data, we show that performance scales consi
Multiple Sclerosis is a chronic autoimmune disorder characterized by the degradation of the myelin sheath in the central nervous system, leading to neurological impairments. In this work, we analyze a reaction-diffusion model derived from kinetic theory to study the formation of demyelinating lesions. We perform a Turing instability analysis and a weakly nonlinear analysis to investigate different spatial patterns that may emerge. Our study examines how key parameters, including the squeezing probability of immune cells and the chemotactic response, impact pattern formation. Numerical simulations confirm the analytical results, revealing the emergence of distinct spatial structures.
Cytotoxic T lymphocytes eliminate infected or malignant cells, safeguarding surrounding tissues. Although experimental and systems-immunology studies have cataloged many molecular and cellular actors involved in an immune response, the design principles governing how the speed and magnitude of T-cell responses emerge from cellular decision-making remain elusive. Here, we recast the T-cell response as a feedback-controlled program, wherein the rates of activation, proliferation, differentiation and death are regulated through antigenic, pro- and anti-inflammatory cues. By exploring a broad class of feedback-controller designs as potential immune programs, we demonstrate how the speed and magnitude of T-cell responses emerge from optimizing signal-feedback to protect against diverse infection settings. We recover an inherent trade-off: infection clearance at the cost of immunopathology. We show how this trade-off is encoded into the logic of T-cell responses by hierarchical sensitivity to different immune signals. Notably, we find that designs that balance harm from acute infections and autoimmunity produce immune responses consistent with experimentally observed patterns of T-cell e
MacroB12 interference presents a significant challenge in the diagnosis of Vitamin B12 status, potentially masking true deficiency. To establish robust predictors and quantify the utility of pre-polyethylene glycol (PEG) B12 levels, we conducted a retrospective analysis of 875 individuals with hypercobalaminaemia (greater than 1000 pg/mL), using multiple imputation to handle missing data. Multivariable regression modelling revealed that MacroB12 positivity (less than 30% PEG recovery) was independently associated with a profile suggestive of autoimmunity, including older age (adjusted odds ratio [aOR] 1.03 per year, P<0.001) and a Rheumatologic/Autoimmune diagnosis (aOR 2.96, P=0.01). We also identified a counterintuitive haematological signature, with higher haemoglobin (aOR 1.14, P=0.02) and Mean Corpuscular Volume (aOR 1.04, P=0.01) predicting MacroB12. Receiver Operating Characteristic (ROC) analysis of pre-PEG B12 concentration yielded only moderate discriminatory power (Area Under the Curve [AUC] = 0.744, 95% CI 0.707-0.782). The optimal threshold of 1584.0 pg/mL (sensitivity 71.3%, specificity 69.7%) serves to stratify clinical suspicion, but its performance confirms that
To ensure faithful information transmission, cells utilize nonequilibrium drives to reduce errors. Kinetic proofreading is a classic mechanism known to sharpen ligand discrimination by T lymphocytes. It remains an open question whether the adaptive immune system relies solely on kinetic proofreading to boost fidelity. Here, we suggest an alternative: an enhanced form of mechanical proofreading (MPR) in which adaptive force exertion via dynamic cell-cell contact allows faithful selection of high-affinity B lymphocytes. Using a minimal model validated with experiment, we show that adaptive MPR, characterized by mechanical feedback between force generation and contact formation, enables robust discrimination of receptor quality regardless of ligand quantity. Although MPR generically balances the tradeoffs between speed and fidelity, a negative scaling of contact duration with ligand abundance indicates the presence of feedback. Due to its ability to modulate interactions of distinct ligands that share load at membrane contacts, adaptive MPR can be harnessed to mitigate autoimmunity or enhance multivalent vaccines. Overall, this work suggests a generalization of the proofreading mechan
The transition from static chat bots to autonomous agents--equipped with persistent memory, tool-use protocols, and multi-agent collaboration--has fundamentally expanded the AI threat landscape. Current defense mechanisms, such as perimeter security and training-time alignment, remain external to the agent's active reasoning loop. Consequently, they fall short: a fully aligned agent remains highly vulnerable to runtime hijacking via memory poisoning, tool-chain manipulation, or multi-agent protocol attacks. To address this critical gap, we introduce the Agent-Native Immune System (ANIS), the first biologically inspired, endogenous defense architecture embedded directly within the agent's cognitive loop. Our framework presents four primary contributions. First, we design a six-layer Immune Tower (L0-L5), distinctly incorporating Barrier Immunity (L1) as a non-cognitive, physical-and-logical isolation layer. Second, we establish a unified taxonomy of Agent Viruses and Agent Vaccines, formalizing the critical distinction between superficial non-parametric defenses and robust parametric vaccines. Third, we conceptualize the Harness Triad--Meta, Self, and Auto--a self-monitoring, meta-c
Replicability is a lynchpin for credible discoveries. The partial conjunction (PC) p-value, which combines individual base p-values from multiple similar studies, can gauge whether a feature of interest exhibits replicated signals across studies. However, when a large set of features are examined as in high-throughput experiments, testing for their replicated signals simultaneously can pose a very underpowered problem, due to both the multiplicity burden and inherent limitations of PC $p$-values. This power deficiency is markedly severe when replication is demanded for all studies under consideration, which is nonetheless the most natural and appealing benchmark for scientific generalizability a practitioner may request. We propose ParFilter, a general framework that marries the ideas of filtering and covariate-adaptiveness to power up large-scale testing for replicated signals as described above. It reduces the multiplicity burden by partitioning studies into smaller groups and borrowing the cross-group information to filter out unpromising features. Moreover, harnessing side information offered by auxiliary covariates whenever they are available, it can train informative hypothes
Demyelinating disorders of the central nervous system may have multiple causes, the most common are infections, autoimmune responses, genetic or vascular etiology. Demyelination lesions are characterized by areas were the myelin sheath of the nerve fibers are broken or destroyed. Among autoimmune disorders, Multiple Sclerosis (MS) is the most well-known Among these disorders, Multiple Sclerosis (MS) is the most well-known and aggressive form. Acute Disseminated Encephalomyelitis (ADEM) is another type of demyelinating disease, typically with a better prognosis. Magnetic Resonance Imaging (MRI) is widely used for diagnosing and monitoring disease progression by detecting lesions. While both adults and children can be affected, there is a significant lack of publicly available datasets for pediatric cases and demyelinating disorders beyond MS. This study introduces, for the first time, a publicly available pediatric dataset for demyelinating lesion segmentation. The dataset comprises MRI scans from 13 pediatric patients diagnosed with demyelinating disorders, including 3 with ADEM. In addition to lesion segmentation masks, the dataset includes extensive patient metadata, such as diag
Prokaryotes have evolved various mechanisms to counter viruses, which in their turn developed numerous strategies to avoid defenses of the hosts. Dozens of such defense and counter-defense mechanisms have recently been discovered, yet the number of such systems held by a given virus or its host is limited. Here, we present numerical and theoretical arguments for the existence of the maximal number of ecologically and evolutionary sustainable defense and counter-defense systems maintained by both sides at any time of the never-ending evolutionary arms race. We find that the number of such systems is of the order of 10 for a broad range of assumptions about the costs and benefits of defense and counter-defense mechanisms and their specificity. This optimum in the number of defense and counter-defense systems appears as a result of a compromise between the metabolic and autoimmune costs of adding a new layer of defense and the benefits it conveys.
Recent advances in artificial intelligence (AI) and multimodal data collection are revolutionizing dermatology. Generative AI and machine learning approaches offer opportunities to enhance the diagnosis and treatment of inflammatory skin diseases, including atopic dermatitis, psoriasis, hidradenitis suppurativa, and autoimmune connective tissue disease. This review examines the current landscape of AI applications for inflammatory skin diseases and explores how generative AI and machine learning methods can advance the field through deep phenotyping, disease heterogeneity characterization, drug development, personalized medicine, and clinical care. We discuss the promises and challenges of these technologies and present a vision for their integration into clinical practice.
T cell receptor (TCR) repertoires encode critical immunological signatures for autoimmune diseases, yet their clinical application remains limited by sequence sparsity and low witness rates. We developed EAMil, a multi-instance deep learning framework that leverages TCR sequencing data to diagnose systemic lupus erythematosus (SLE) and rheumatoid arthritis (RA) with exceptional accuracy. By integrating PrimeSeq feature extraction with ESMonehot encoding and enhanced gate attention mechanisms, our model achieved state-of-the-art performance with AUCs of 98.95% for SLE and 97.76% for RA. EAMil successfully identified disease-associated genes with over 90% concordance with established differential analyses and effectively distinguished disease-specific TCR genes. The model demonstrated robustness in classifying multiple disease categories, utilizing the SLEDAI score to stratify SLE patients by disease severity as well as to diagnose the site of damage in SLE patients, and effectively controlling for confounding factors such as age and gender. This interpretable framework for immune receptor analysis provides new insights for autoimmune disease detection and classification with broad p