Marcel J. T. Reinders

The paper evaluates antibody design methods using immunogenicity labels and finds that format-based approaches outperform sequence-based predictors, but label inaccuracies limit progress.

This study investigates when combining mechanistic and machine learning models improves predictions of gene regulatory network dynamics using synthetic data.

This study compares endogenous viral elements in Alzheimer's Disease patients and cognitively healthy centenarians using long-read sequencing, finding higher EVE burden in AD patients.

The paper introduces SACCELERATOR, a framework for spatially aware clustering that improves benchmarking by incorporating expert feedback and consensus analysis across diverse datasets.

The paper uses long-read sequencing to uncover structural complexity in Alzheimer's disease GWAS loci, showing that lead SNPs often miss key regulatory variants.

The paper develops a data-driven framework to detect cancer using cfDNA fragmentation patterns at TSSs, outperforming existing methods in distinguishing cancer from healthy samples.

The paper identifies and characterizes distinct B cell subsets in the human thymus, revealing their development, differentiation, and class switching capabilities.

The paper shows that drug response predictors trained on perturbation response model-generated profiles outperform those trained on baseline or pre-treatment profiles.

This paper introduces a discrete diffusion model for protein sequence design using all-atom representations, showing improved novelty and diversity compared to amino acid-based models.

This study evaluates the impact of Gene Ontology priors in biologically-informed neural networks using RNA-seq data, finding limited performance improvement and interpretability beyond predefined gene sets.

This study constructs a polygenic protective score for Alzheimer's disease and finds stronger genetic protection against AD in centenarians compared to non-longevity individuals.

The paper identifies six HLA alleles associated with reduced likelihood of becoming a cognitively healthy centenarian, revealing pleiotropic effects on disease susceptibility and longevity.

The paper identifies genetic variants associated with protein abundance and altered protein-protein correlations in the Gyrus Temporalis Medius in Alzheimer's Disease patients.

This paper compares federated and centralized learning through experiments, finding similar performance and federated learning's robustness to data imbalances and high dimensionality.

This paper introduces scAGG, a model for classifying Alzheimer's disease from single-nucleus RNA-seq data by aggregating single-cell embeddings, and finds it accurately classifies AD and healthy controls with disease severity ordering in the latent space.

BADDADAN models gene module dynamics under stress using machine learning and ODEs, demonstrating its application on A. thaliana datasets to recover known insights and propose new hypotheses.

The paper introduces otter and TREAT for accurate tandem repeat genotyping across long-read sequencing platforms, showing improved accuracy and identifying pathogenic expansions and disease associations.

This study examines how a lifestyle intervention affects the blood transcriptomic and metabolomic responses to a nutrient challenge in elderly individuals.

The paper analyzes exome data from 22,319 individuals to identify rare CNVs and 22q11.21 dosage as risk factors for Alzheimer's disease.

The paper introduces gsQTL, a method that links genetic risk variants to gene sets by analyzing shared variability, showing improved performance over traditional approaches.

This paper designs antibody sequences and structures using a diffusion model guided by developability properties like solubility and folding stability.

The paper predicts cell-type-specific exon inclusion in human brain cells using RBP binding sites and genomic sequence, revealing more complex splicing mechanisms in neurons than glia.

This paper introduces a decision tree to improve the identification of dementia patients with pathogenic genetic variants by optimizing clinical selection criteria for genetic testing.

This study compares linear and non-linear joint embedding methods for bulk and single-cell multi-omics, finding non-linear methods excel in imputation and joint embeddings improve performance in supervised tasks.

This paper uses transformer models to predict T-cell receptor epitope specificity, finding that modified ProtXLNet achieves the highest accuracy among tested models.

This study benchmarks variational autoencoders on cancer transcriptomics data to evaluate hyperparameter effects and latent space biological relevance.

The paper identifies an AluYb8 retrotransposon in a TMEM106B risk haplotype associated with neurodegeneration, linking it to age-related dysregulation and population-specific disease risk.

The paper identifies high-priority missense variants in SORL1 associated with increased Alzheimer's Disease risk and earlier onset, suggesting clinical counseling considerations for carriers.