Markus Heinonen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Markus Heinonen has a total of 58 co-authors. Here are the top co-authors by number of shared publications:
- Emmi Jokinen
Hematology Research Unit Helsinki, University of Helsinki and Helsinki University Hospital, Helsinki, Finland
4 shared publications - Jani Huuhtanen
Helsinki University Hospital Comprehensive Cancer Center,Translational Immunology Research Program and Department of Clinical Chemistry and Hematology
4 shared publications - Satu Mustjoki
Translational Immunology Research program and Department of Clinical Chemistry and Hematology, University of Helsinki, Helsinki, Finland
4 shared publications - Samuel Kaski
Department of Computer Science, Aalto University, Espoo, Finland
3 shared publications - Eva Hellström-Lindberg
Department of Medicine Huddinge, Center for Hematology and Regenerative Medicine, Karolinska Institute, Huddinge, Sweden
2 shared publications - Freja Ebeling
Division of Hematology, Helsinki University Hospital, Comprehensive Cancer Center, Helsinki, Finland
2 shared publications
Latest Publications
- Open Access
Deep latent variable modelling reveals clinically significant subgroups among transfusion recipients
A bstract: Background: Transfusion recipients are a heterogeneous group of patients, yet the identification of these groups has traditionally relied on human-driven univariate analyses and domain knowledge instead of analyzing multivariate characteristics of individuals. Electronic health records (EHR) combined with unsupervised machine learning enables robust, data-driven way for phenotyping patient populations, providing finer-grained view on subgroup characteristics. Materials and Methods: We introduce an extension to the Variational Autoencoder (VAE) framework and apply the model to EHR data of 19,629 adult transfusion recipients. The latent representation of VAEs approximates a low-dimensional manifold of input data, in which patients with similar characteristics are embedded close to one another. The model integrates clustering via a Gaussian Mixture Model (GMM) prior to identify clinically relevant patient subgroups from diagnosis codes, laboratory values and demographics, while simultaneously classifying the type of transfused products. The final clusters are derived with a modified consensus clustering approach. Results: We identified six patient groups with distinct diagnosis, laboratory, demographic, and transfusion profiles. These novel clusters provide a refined characterization of transfusion-related phenotypes, revealing more detailed distinctions among patient subgroups. Our model achieved moderate classification accuracy, with AUROC of 0.879, 0.806 and 0.861, and PR-AUC of 0.448, 0.357 and 0.492 for red blood cells (RBC), plasma and platelets, respectively. Clustering accuracy remains consistent across training and test sets. Conclusions: Data-driven phenotyping of transfusion recipients revealed previously unexplored patient phenotypes differing in multiple characteristics. The model helps to understand the heterogeneous nature of patients requiring transfusion and provides insights on how different blood product profiles shift cluster assignments. These findings underscore the utility of latent variable modelling for population characterization and suggest potential applications in optimizing transfusion strategies as well as blood supply chain management. Validation in external cohorts remains unestablished.
Open AccessDeep latent variable modelling reveals clinically significant subgroups among transfusion recipients
A bstract: Background: Transfusion recipients are a heterogeneous group of patients, yet the identification of these groups has traditionally relied on human-driven univariate analyses and domain knowledge instead of analyzing multivariate characteristics of individuals. Electronic health records (EHR) combined with unsupervised machine learning enables robust, data-driven way for phenotyping patient populations, providing finer-grained view on subgroup characteristics. Materials and Methods: We introduce an extension to the Variational Autoencoder (VAE) framework and apply the model to EHR data of 19,629 adult transfusion recipients. The latent representation of VAEs approximates a low-dimensional manifold of input data, in which patients with similar characteristics are embedded close to one another. The model integrates clustering via a Gaussian Mixture Model (GMM) prior to identify clinically relevant patient subgroups from diagnosis codes, laboratory values and demographics, while simultaneously classifying the type of transfused products. The final clusters are derived with a modified consensus clustering approach. Results: We identified six patient groups with distinct diagnosis, laboratory, demographic, and transfusion profiles. These novel clusters provide a refined characterization of transfusion-related phenotypes, revealing more detailed distinctions among patient subgroups. Our model achieved moderate classification accuracy, with AUROC of 0.879, 0.806 and 0.861, and PR-AUC of 0.448, 0.357 and 0.492 for red blood cells (RBC), plasma and platelets, respectively. Clustering accuracy remains consistent across training and test sets. Conclusions: Data-driven phenotyping of transfusion recipients revealed previously unexplored patient phenotypes differing in multiple characteristics. The model helps to understand the heterogeneous nature of patients requiring transfusion and provides insights on how different blood product profiles shift cluster assignments. These findings underscore the utility of latent variable modelling for population characterization and suggest potential applications in optimizing transfusion strategies as well as blood supply chain management. Validation in external cohorts remains unestablished.
- Open Access
- 23 May 2025
- Published by IOP Publishing in Machine Learning: Science and Technology
- Vol. 6 (2) , 025042
Multi-target property prediction and optimization using latent spaces of generative model
Multi-target property prediction has the potential to improve generalization by exploiting the positive transfer between targets. Molecular generative models utilize independent single-target property prediction networks to discover novel molecules. We propose using multi-target networks to jointly predict several molecular properties and learn better representations by exploiting auxiliary information. Our multi-target model shows improvement in prediction accuracy on the test set. We additionally present results demonstrating promising performance in property prediction in these generative models does not translate to optimization. More specifically, random exploration is competitive with gradient-based strategies and better methods are needed.
- 23 May 2025
- Published by IOP Publishing in Machine Learning: Science and Technology
- Vol. 6 (2) , 025042
Open AccessMulti-target property prediction and optimization using latent spaces of generative model
Multi-target property prediction has the potential to improve generalization by exploiting the positive transfer between targets. Molecular generative models utilize independent single-target property prediction networks to discover novel molecules. We propose using multi-target networks to jointly predict several molecular properties and learn better representations by exploiting auxiliary information. Our multi-target model shows improvement in prediction accuracy on the test set. We additionally present results demonstrating promising performance in property prediction in these generative models does not translate to optimization. More specifically, random exploration is competitive with gradient-based strategies and better methods are needed.
- 26 February 2025
- Published by American Association for the Advancement of Science (AAAS) in Science Translational Medicine
- Vol. 17 (787) , eadl6758
Single-cell analysis of aplastic anemia reveals a convergence of NK and NK-like CD8 + T cells with a disease-associated TCR signature
Immune aplastic anemia (AA) is a life-threatening bone marrow failure disorder driven by an autoimmune T cell attack against hematopoietic stem and progenitor cells (HSPCs). However, the exact autoantigen targets and role of other immune cells in the pathogenesis of AA are unknown. Here, we analyzed a cohort of 218 patients with AA using single-cell RNA and T cell receptor (TCR) αβ sequencing, TCRβ sequencing, flow cytometry, and plasma cytokine profiling. We identified natural killer (NK) cells and CD8 + terminally differentiated effector T (T EMRA ) cells expressing NK receptors with AA-associated TCRβ motifs as the most dysregulated immune cell populations in AA bone marrow. Functional coculture experiments using primary HSPCs and immune cells showed that NK cells cannot kill HSPCs alone but may sensitize HSPCs to CD8 + T cell–mediated killing through production of interferons. Furthermore, HSPCs induced activation of T cell clones with CD8 + T EMRA NK-like phenotype in coculture. Our results reveal a convergent phenotype of innate and adaptive immune cells that may drive AA.
- 26 February 2025
- Published by American Association for the Advancement of Science (AAAS) in Science Translational Medicine
- Vol. 17 (787) , eadl6758
Single-cell analysis of aplastic anemia reveals a convergence of NK and NK-like CD8 + T cells with a disease-associated TCR signature
Immune aplastic anemia (AA) is a life-threatening bone marrow failure disorder driven by an autoimmune T cell attack against hematopoietic stem and progenitor cells (HSPCs). However, the exact autoantigen targets and role of other immune cells in the pathogenesis of AA are unknown. Here, we analyzed a cohort of 218 patients with AA using single-cell RNA and T cell receptor (TCR) αβ sequencing, TCRβ sequencing, flow cytometry, and plasma cytokine profiling. We identified natural killer (NK) cells and CD8 + terminally differentiated effector T (T EMRA ) cells expressing NK receptors with AA-associated TCRβ motifs as the most dysregulated immune cell populations in AA bone marrow. Functional coculture experiments using primary HSPCs and immune cells showed that NK cells cannot kill HSPCs alone but may sensitize HSPCs to CD8 + T cell–mediated killing through production of interferons. Furthermore, HSPCs induced activation of T cell clones with CD8 + T EMRA NK-like phenotype in coculture. Our results reveal a convergent phenotype of innate and adaptive immune cells that may drive AA.
- Open Access
- 20 September 2024
- Published by Springer Nature
- p. 82-97
Balancing Imbalanced Toxicity Models: Using MolBERT with Focal Loss
Drug-induced liver injury (DILI) presents a multifaceted challenge, influenced by interconnected biological mechanisms. Current DILI datasets are characterized by small sizes and high imbalance, posing difficulties in learning robust representations and accurate modeling. To address these challenges, we trained a multi-modal multi-task model integrating preclinical histopathologies, biochemistry (blood markers), and clinical DILI-related adverse drug reactions (ADRs). Leveraging pretrained BERT models, we extracted representations covering a broad chemical space, facilitating robust learning in both frozen and fine-tuned settings. To address imbalanced data, we explored weighted Binary Cross-Entropy (w-BCE) and weighted Focal Loss (w-FL) . Our results demonstrate that the frozen BERT model consistently enhances performance across all metrics and modalities with weighted loss functions compared to their non-weighted counterparts. However, the efficacy of fine-tuning BERT varies across modalities, yielding inconclusive results. In summary, the incorporation of BERT features with weighted loss functions demonstrates advantages, while the efficacy of fine-tuning remains uncertain.
- 20 September 2024
- Published by Springer Nature
- p. 82-97
Open AccessBalancing Imbalanced Toxicity Models: Using MolBERT with Focal Loss
Drug-induced liver injury (DILI) presents a multifaceted challenge, influenced by interconnected biological mechanisms. Current DILI datasets are characterized by small sizes and high imbalance, posing difficulties in learning robust representations and accurate modeling. To address these challenges, we trained a multi-modal multi-task model integrating preclinical histopathologies, biochemistry (blood markers), and clinical DILI-related adverse drug reactions (ADRs). Leveraging pretrained BERT models, we extracted representations covering a broad chemical space, facilitating robust learning in both frozen and fine-tuned settings. To address imbalanced data, we explored weighted Binary Cross-Entropy (w-BCE) and weighted Focal Loss (w-FL) . Our results demonstrate that the frozen BERT model consistently enhances performance across all metrics and modalities with weighted loss functions compared to their non-weighted counterparts. However, the efficacy of fine-tuning BERT varies across modalities, yielding inconclusive results. In summary, the incorporation of BERT features with weighted loss functions demonstrates advantages, while the efficacy of fine-tuning remains uncertain.
- Open Access
Modular Flows: Differential Molecular Generation
Generating new molecules is fundamental to advancing critical applications such as drug discovery and material synthesis. Flows can generate molecules effectively by inverting the encoding process, however, existing flow models either require artifactual dequantization or specific node/edge orderings, lack desiderata such as permutation invariance, or induce discrepancy between the encoding and the decoding steps that necessitates post hoc validity correction. We circumvent these issues with novel continuous normalizing E(3)-equivariant flows, based on a system of node ODEs coupled as a graph PDE, that repeatedly reconcile locally toward globally aligned densities. Our models can be cast as message-passing temporal networks, and result in superlative performance on the tasks of density estimation and molecular generation. In particular, our generated samples achieve state-of-the-art on both the standard QM9 and ZINC250K benchmarks.
Open AccessModular Flows: Differential Molecular Generation
Generating new molecules is fundamental to advancing critical applications such as drug discovery and material synthesis. Flows can generate molecules effectively by inverting the encoding process, however, existing flow models either require artifactual dequantization or specific node/edge orderings, lack desiderata such as permutation invariance, or induce discrepancy between the encoding and the decoding steps that necessitates post hoc validity correction. We circumvent these issues with novel continuous normalizing E(3)-equivariant flows, based on a system of node ODEs coupled as a graph PDE, that repeatedly reconcile locally toward globally aligned densities. Our models can be cast as message-passing temporal networks, and result in superlative performance on the tasks of density estimation and molecular generation. In particular, our generated samples achieve state-of-the-art on both the standard QM9 and ZINC250K benchmarks.
- Open Access
Tackling covariate shift with node-based Bayesian neural networks
Bayesian neural networks (BNNs) promise improved generalization under covariate shift by providing principled probabilistic representations of epistemic uncertainty. However, weight-based BNNs often struggle with high computational complexity of large-scale architectures and datasets. Node-based BNNs have recently been introduced as scalable alternatives, which induce epistemic uncertainty by multiplying each hidden node with latent random variables, while learning a point-estimate of the weights. In this paper, we interpret these latent noise variables as implicit representations of simple and domain-agnostic data perturbations during training, producing BNNs that perform well under covariate shift due to input corruptions. We observe that the diversity of the implicit corruptions depends on the entropy of the latent variables, and propose a straightforward approach to increase the entropy of these variables during training. We evaluate the method on out-of-distribution image classification benchmarks, and show improved uncertainty estimation of node-based BNNs under covariate shift due to input perturbations. As a side effect, the method also provides robustness against noisy training labels.
Open AccessTackling covariate shift with node-based Bayesian neural networks
Bayesian neural networks (BNNs) promise improved generalization under covariate shift by providing principled probabilistic representations of epistemic uncertainty. However, weight-based BNNs often struggle with high computational complexity of large-scale architectures and datasets. Node-based BNNs have recently been introduced as scalable alternatives, which induce epistemic uncertainty by multiplying each hidden node with latent random variables, while learning a point-estimate of the weights. In this paper, we interpret these latent noise variables as implicit representations of simple and domain-agnostic data perturbations during training, producing BNNs that perform well under covariate shift due to input corruptions. We observe that the diversity of the implicit corruptions depends on the entropy of the latent variables, and propose a straightforward approach to increase the entropy of these variables during training. We evaluate the method on out-of-distribution image classification benchmarks, and show improved uncertainty estimation of node-based BNNs under covariate shift due to input perturbations. As a side effect, the method also provides robustness against noisy training labels.
- Open Access
Predicting recognition between T cell receptors and epitopes using contextualized motifs
We introduce TCRconv, a deep learning model for predicting recognition between T-cell receptors and epitopes. TCRconv uses a deep protein language model and convolutions to extract contextualized motifs and provides state-of-the-art TCR-epitope prediction accuracy. Using TCR repertoires from COVID-19 patients, we demonstrate that TCRconv can provide insight into T-cell dynamics and phenotypes during the disease.
Open AccessPredicting recognition between T cell receptors and epitopes using contextualized motifs
We introduce TCRconv, a deep learning model for predicting recognition between T-cell receptors and epitopes. TCRconv uses a deep protein language model and convolutions to extract contextualized motifs and provides state-of-the-art TCR-epitope prediction accuracy. Using TCR repertoires from COVID-19 patients, we demonstrate that TCRconv can provide insight into T-cell dynamics and phenotypes during the disease.
- Open Access
- 25 March 2021
- Published by Public Library of Science (PLoS) in PLOS Computational Biology
- Vol. 17 (3) , e1008814
Predicting recognition between T cell receptors and epitopes with TCRGP
Adaptive immune system uses T cell receptors (TCRs) to recognize pathogens and to consequently initiate immune responses. TCRs can be sequenced from individuals and methods analyzing the specificity of the TCRs can help us better understand individuals’ immune status in different disorders. For this task, we have developed TCRGP, a novel Gaussian process method that predicts if TCRs recognize specified epitopes. TCRGP can utilize the amino acid sequences of the complementarity determining regions (CDRs) from TCRα and TCRβ chains and learn which CDRs are important in recognizing different epitopes. Our comprehensive evaluation with epitope-specific TCR sequencing data shows that TCRGP achieves on average higher prediction accuracy in terms of AUROC score than existing state-of-the-art methods in epitope-specificity predictions. We also propose a novel analysis approach for combined single-cell RNA and TCRαβ (scRNA+TCRαβ) sequencing data by quantifying epitope-specific TCRs with TCRGP and identify HBV-epitope specific T cells and their transcriptomic states in hepatocellular carcinoma patients.
- 25 March 2021
- Published by Public Library of Science (PLoS) in PLOS Computational Biology
- Vol. 17 (3) , e1008814
Open AccessPredicting recognition between T cell receptors and epitopes with TCRGP
Adaptive immune system uses T cell receptors (TCRs) to recognize pathogens and to consequently initiate immune responses. TCRs can be sequenced from individuals and methods analyzing the specificity of the TCRs can help us better understand individuals’ immune status in different disorders. For this task, we have developed TCRGP, a novel Gaussian process method that predicts if TCRs recognize specified epitopes. TCRGP can utilize the amino acid sequences of the complementarity determining regions (CDRs) from TCRα and TCRβ chains and learn which CDRs are important in recognizing different epitopes. Our comprehensive evaluation with epitope-specific TCR sequencing data shows that TCRGP achieves on average higher prediction accuracy in terms of AUROC score than existing state-of-the-art methods in epitope-specificity predictions. We also propose a novel analysis approach for combined single-cell RNA and TCRαβ (scRNA+TCRαβ) sequencing data by quantifying epitope-specific TCRs with TCRGP and identify HBV-epitope specific T cells and their transcriptomic states in hepatocellular carcinoma patients.
- Open Access
- 29 November 2018
- Published by American Society of Hematology in Blood
- Vol. 132 (Supplement) , 537
Next-Generation Sequencing Reveals a T Cell Receptor Signature Characteristic of Patients with Aplastic Anemia
Background. Acquired aplastic anemia (AA) is a bone marrow failure syndrome, in which patients' hematopoietic stem cells are destroyed, resulting in pancytopenia. The exact mechanism and biological process leading to AA remain largely unknown. Bone marrow destruction is perceived as an immune-mediated process, which is supported by elevated cytotoxic T lymphocyte (CTL) counts, responsiveness to immunosuppressive therapy and skewed CTL T cell receptor (TCR) repertoire. Although there is a well-established role of T cells in the pathology of AA, the putative antigen behind the autoimmune response, and thus, the TCRs recognising the antigen are still unrevealed. Methods. Our cohort comprised of 130 samples, consisting of bone marrow and/or peripheral blood samples from AA patients (n=52) from diagnosis and/or follow-up phases and from healthy controls (n=27). We performed TCRβ sequencing (immunoSEQ, Adaptive Biotechnologies) on sorted AA and healthy CTLs (n=25 and n=27) or AA mononuclear cells (MNCs, n=27). To gain in-depth understanding of the TCR-repertoire, we built two novel analysis methods: 1) an unsupervised clustering method to characterize epitope-specific T cells based on the amino acid -level similarities of TCRs and 2) a probability based classifier to identify AA based on TCR data in the CTL (discovery) and MNC (validation) cohorts. Results. To fully appreciate the complex nature of AA pathology, we divided the cytotoxic T cell response in two distinct categories, private and public response. Private response comprises T cell clones found only in individual patients, and this compartment could explain the variation in treatment responses and disease severity across patients. In the CTL TCR repertoire, we identified patient exclusive expanded T cell clones (>1% frequency in TCR repertoire, n = 317) and treatment responding clones (n = 364). Clustering analysis revealed that there is significant amino acid -level similarity between the TCRs of the private clones. We found several epitope-specific TCR clusters that were associated with different treatment responses, disease severities and HLA-DR15 risk-allele positivity. Interestingly, we discovered that the public response (CTL clones that are statistically enriched in AA patients compared to healthy controls) could discriminate AA from healthy TCR repertoire. Based on these publicly enriched TCRs, we built a classifier which could identify AA CTL TCR repertoire from healthy controls with 97% accuracy (F-score, revieved from leave-one-out cross-validation). We tested our classifier with the validation cohort. The accuracy to diagnose AA based on the TCR repertoire data in the MNC cohort was 0.72. Furthermore, the public clones that differentiated best the AA cases from healthy controls showed statistically significant peptide similarity. These public TCRs occurred also in healthy controls, but with smaller frequencies. When combining the private and public response TCRs, we discovered that some of AA patients' most expanded and treatment responding clones clustered in the same putative epitope-specific clusters with the public TCRs, showing an interesting intersection between public and private signatures. In addition, the comparison of the interesting private and public TCRs against databases of TCR sequences of known antigen specificities hinted that some of these clones may originally target known viral species (CMV, EBV and Influenza A), suggesting a role of these common pathogens in the development of AA. Discussion. CTLTCR repertoire analysis of AA patients revealed a TCR signature that was typical of AA patients, but varied between different patients, and it was validated with an independent dataset. Thus, we could design a TCR-based framework which could identify AA patients based on their TCR repertoire, independently of sample type. A future application for our classifier could be distinguishing AA from other AA-like diseases, like hypoplastic myelodysplastic syndrome. Furthermore, we found groups of TCRs that look similar on amino acid level, and hence these clones may target the same epitope. These TCR clusters were associated with clinical features. Amino acid similarity -based TCR signatures or TCR classifiers have not previously been published for AA or any other autoimmune diseases, and thus our pioneering tools could be utilized to study pathogenesis of other T cell mediated diseases as well.
- 29 November 2018
- Published by American Society of Hematology in Blood
- Vol. 132 (Supplement) , 537
Open AccessNext-Generation Sequencing Reveals a T Cell Receptor Signature Characteristic of Patients with Aplastic Anemia
Background. Acquired aplastic anemia (AA) is a bone marrow failure syndrome, in which patients' hematopoietic stem cells are destroyed, resulting in pancytopenia. The exact mechanism and biological process leading to AA remain largely unknown. Bone marrow destruction is perceived as an immune-mediated process, which is supported by elevated cytotoxic T lymphocyte (CTL) counts, responsiveness to immunosuppressive therapy and skewed CTL T cell receptor (TCR) repertoire. Although there is a well-established role of T cells in the pathology of AA, the putative antigen behind the autoimmune response, and thus, the TCRs recognising the antigen are still unrevealed. Methods. Our cohort comprised of 130 samples, consisting of bone marrow and/or peripheral blood samples from AA patients (n=52) from diagnosis and/or follow-up phases and from healthy controls (n=27). We performed TCRβ sequencing (immunoSEQ, Adaptive Biotechnologies) on sorted AA and healthy CTLs (n=25 and n=27) or AA mononuclear cells (MNCs, n=27). To gain in-depth understanding of the TCR-repertoire, we built two novel analysis methods: 1) an unsupervised clustering method to characterize epitope-specific T cells based on the amino acid -level similarities of TCRs and 2) a probability based classifier to identify AA based on TCR data in the CTL (discovery) and MNC (validation) cohorts. Results. To fully appreciate the complex nature of AA pathology, we divided the cytotoxic T cell response in two distinct categories, private and public response. Private response comprises T cell clones found only in individual patients, and this compartment could explain the variation in treatment responses and disease severity across patients. In the CTL TCR repertoire, we identified patient exclusive expanded T cell clones (>1% frequency in TCR repertoire, n = 317) and treatment responding clones (n = 364). Clustering analysis revealed that there is significant amino acid -level similarity between the TCRs of the private clones. We found several epitope-specific TCR clusters that were associated with different treatment responses, disease severities and HLA-DR15 risk-allele positivity. Interestingly, we discovered that the public response (CTL clones that are statistically enriched in AA patients compared to healthy controls) could discriminate AA from healthy TCR repertoire. Based on these publicly enriched TCRs, we built a classifier which could identify AA CTL TCR repertoire from healthy controls with 97% accuracy (F-score, revieved from leave-one-out cross-validation). We tested our classifier with the validation cohort. The accuracy to diagnose AA based on the TCR repertoire data in the MNC cohort was 0.72. Furthermore, the public clones that differentiated best the AA cases from healthy controls showed statistically significant peptide similarity. These public TCRs occurred also in healthy controls, but with smaller frequencies. When combining the private and public response TCRs, we discovered that some of AA patients' most expanded and treatment responding clones clustered in the same putative epitope-specific clusters with the public TCRs, showing an interesting intersection between public and private signatures. In addition, the comparison of the interesting private and public TCRs against databases of TCR sequences of known antigen specificities hinted that some of these clones may originally target known viral species (CMV, EBV and Influenza A), suggesting a role of these common pathogens in the development of AA. Discussion. CTLTCR repertoire analysis of AA patients revealed a TCR signature that was typical of AA patients, but varied between different patients, and it was validated with an independent dataset. Thus, we could design a TCR-based framework which could identify AA patients based on their TCR repertoire, independently of sample type. A future application for our classifier could be distinguishing AA from other AA-like diseases, like hypoplastic myelodysplastic syndrome. Furthermore, we found groups of TCRs that look similar on amino acid level, and hence these clones may target the same epitope. These TCR clusters were associated with clinical features. Amino acid similarity -based TCR signatures or TCR classifiers have not previously been published for AA or any other autoimmune diseases, and thus our pioneering tools could be utilized to study pathogenesis of other T cell mediated diseases as well.