Christian Herder

LEOPARD completes missing views in multi-timepoint omics data by disentangling content and temporal representations, outperforming existing methods on benchmark datasets.

The paper develops an interpretable multimodal machine learning framework to predict distal sensorimotor polyneuropathy using clinical and molecular data, identifying key biomarkers and improving prediction accuracy with additional molecular modalities.

The paper identifies pQTLs for 91 plasma proteins, finds novel associations, and uses MR to uncover causal roles in immune-mediated diseases, suggesting new therapeutic targets.

The paper uses proteogenomic integration to identify CXCL10 as a mediator of IL-6 signaling effects on atherosclerosis, showing its association with cardiovascular disease risk and lesion characteristics.

The paper reconstructs regulatory networks underlying trans-QTL hotspots using human multi-omics data and biological priors, identifying novel networks related to schizophrenia and lean body mass.

This study identifies nine novel DNA methylation sites associated with fasting glucose and insulin levels by integrating cross-omics data from multiple cohorts.

This study fine-maps type 2 diabetes loci to single-variant resolution using high-density imputation and islet-specific epigenome maps, identifying 73 variants with >80% posterior probability of association.

The paper introduces a method for cell-specific eQTL analysis without cell sorting, showing that Crohn's disease-associated SNPs mainly affect neutrophil gene expression in whole blood data.