Holger Fröhlich

The paper introduces a system for predicting tissue-specific E3 ligases for targeted protein degradation using a BERT-based model and integrated biological data.

This study predicts Parkinson's dementia using machine learning and multimodal data, finding genetic predisposition and comorbidities as key factors.

This paper introduces MultiGML, a multimodal graph machine learning approach that predicts adverse drug events by integrating diverse data sources and provides explainable biological insights.

This paper develops an AI model to simulate CD33 down-expression effects on cognitive impairment in Alzheimer's Disease, validated with cell line data.

This paper introduces a deep learning method called STEP to predict virus-host protein-protein interactions using protein sequence embeddings and a Siamese neural network, demonstrating its effectiveness on SARS-CoV-2 and JCV proteins.

The paper presents a drug-target-mechanism graph to identify synergistic drug combinations for SARS-CoV-2 treatment, validated through experimental testing.

GuiltyTargets prioritizes therapeutic targets using deep learning on protein interaction networks, achieving high accuracy in disease-specific target ranking and showing potential for target repositioning.

This paper introduces VAMBN, a variational autoencoder modular Bayesian network for simulating realistic clinical study data while ensuring privacy and enabling counterfactual scenarios.

This study benchmarks the impact of pathway database choice on enrichment analysis and predictive modeling, showing that using multiple or integrative databases improves consistency and performance.