Morten Nielsen

This paper introduces a supervised learning denoising method to identify low-quality labels in TCR specificity data, improving model reliability by filtering noisy examples.

The paper develops a machine learning framework to predict MHC class II binding for canonical and modified peptides by incorporating chemical features, improving immunogenicity risk assessment for non-canonical peptide drugs.

The paper introduces TCRdenoise, a method to denoise TCR-pMHC interaction data by clustering and similarity metrics, improving the quality of training data for predictive models.

The paper shows label noise limits TCR-pMHC specificity prediction and improves performance using AlphaFold3 and denoising.

The paper improves antibody-antigen structure prediction by integrating B-cell epitope prediction tools into Chai-1 modeling, enhancing accuracy and diversity of predicted complexes.

NetMHCIIphosPan predicts HLA class II antigen presentation of phosphorylated peptides using immunopeptidomics data, outperforming existing tools.

IMMREP25 evaluates TCR:pMHC binding prediction for unseen peptides, showing improved performance with structural modeling but still below random guessing.

ITRAP2 improves TCR-pMHC pairing accuracy by reducing noise and cross-contamination in single-cell data, outperforming existing methods on both in-house and public datasets.

The paper characterizes HLA-B*73:01, revealing its unique peptide binding motif, restricted peptidome, and distinct interaction with KIR2DL2, suggesting evolutionary advantages in pathogen resistance.

The paper introduces NetTCR-struc, a GNN-based method to improve TCR-pMHC interaction modeling by enhancing docking quality scoring and structural model selection.

The paper identifies that CDR1 and CDR2 sequence identities are more predictive of epitope similarity than CDR3, and develops BCRMatch to predict epitope binding based on antibody sequences.

This work introduces a deep learning approach using CNNs to discover peptide binding motifs in MHC molecules, showing alignment with experimental motifs and potential applications in vaccine design.

Genesis is a modular protein language model that improves immunogenicity prediction for CD8+ epitopes by integrating pMHC and TCR context, outperforming existing models in precision.

NetTCR 2.2 improves TCR specificity predictions by combining pan- and peptide-specific training, loss-scaling, and sequence similarity integration, achieving state-of-the-art performance on the IMMREP 2022 benchmark.

The paper introduces HLAAssoc-1.0 to infer HLA-DRB3/4/5 alleles from HLA-DRB1 data, improving allele and motif deconvolutions in immunopetidomics and training of binding prediction models.

The paper improves HLA-DQ antigen presentation prediction using machine learning and shows trans-only variants have limited impact on the immunopeptidome.

The paper benchmarks 23 TCR-epitope prediction models, finding that using paired-chain alpha-beta and CDR1/2/V/J data improves performance, and simple distance-based methods can be competitive.

This study benchmarks two denoising methods for TCR-pMHC single-cell data, finding ITRAP outperforms ICON in signal-to-noise ratio and data consistency.

NetAllergen improves allergenicity prediction by integrating MHC-II presentation propensity, outperforming existing methods when sequence similarity to known allergens is low.

BepiPred-3.0 improves B-cell epitope prediction using protein language models, achieving high accuracy for both linear and conformational epitopes.

This study shows that impaired proinsulin handling in β-cells increases MHC-I bound peptides and activates inflammasomes, potentially contributing to autoimmune diabetes.

The paper integrates immunopeptidomics and immunoinformatics to characterize BoLA-DR antigen presentation and identify CD4 T cell epitopes with high predictive accuracy.

This paper improves T cell epitope predictions by combining MHC binding affinities with antigen expression levels, showing better performance than methods using only binding or expression data.

The paper improves MHC II epitope prediction by integrating mass spectrometry data with a new machine learning framework, achieving better performance than existing methods.

NNAlign_MA deconvolves poly-specific MHC peptidome data to accurately characterize binding motifs and improve T cell epitope predictions.

NetSurfP-2.0 improves prediction of protein structural features with higher accuracy and faster runtime than previous versions.

A sequence-only convolutional network trained on T cell receptor (TCR)-peptide pairs for HLA-A*02:01 can single out TCRs that bind a target peptide among many non-binding ones.

The paper describes computational approaches for analyzing immunopeptidomics data, focusing on motif identification, false hit filtering, and machine learning applications for predicting MHC ligands.

NetMHCpan-4.0 improves peptide-MHC class I interaction predictions by integrating binding affinity and eluted ligand data, enhancing identification of naturally processed ligands and T cell epitopes.