Stefan Bonn

BioChemAIgent provides an AI-driven framework for protein modeling and docking in drug discovery, integrating multiple tools into a unified interface.

SCALE identifies multi-scale functional domains in spatial transcriptomics data using neural graph learning and entropy-based search, achieving state-of-the-art performance on simulated and biological datasets.

This study uses spatial transcriptomics to show that immune and renal cell interactions drive glomerular crescent formation in RPGN through PDGF and TGF-β signaling to PECs.

The paper develops an in situ hybridization method to analyze T cell clonality in kidney biopsies and finds spatial clustering of clonally related T cells near antigen-presenting cells.

This paper introduces BioContextAI, an open-source initiative that provides standardized access to validated biomedical knowledge to improve the reliability of AI systems in biomedical research.

This paper introduces NichePCA, a PCA-based method for spatial domain identification that performs competitively with ten state-of-the-art methods across six spatial transcriptomic datasets.

This study shows that neddylation is essential for the development and function of glutamatergic neurons, with Nedd8 deletion causing defects in dendrite complexity and synaptic transmission.

This study identifies podocyte exopher-formation as a novel mechanism for removing antigen/autoantibody aggregates in membranous nephropathy, with potential for non-invasive monitoring.

Seumetry is an R toolkit that integrates scRNA-seq analysis methods with flow and mass cytometry data to enable comprehensive and scalable analysis of high-dimensional cytometry data.

The paper presents a multi-omics dataset for genetic subtypes of frontotemporal dementia to identify molecular mechanisms disrupted in the disease.

DISCERN improves cell clustering and subtype detection by reconstructing missing gene expression, identifying novel T cell types in COVID-19 and predicting disease severity.

This study demonstrates that RNAlater preserves kidney biopsies effectively for single nucleus RNA sequencing, proteomics, and histopathology, but not metabolomics.

The paper performs a multiomic analysis of ALS patients and mouse models to identify sex differences, molecular subclusters, and the MAPK pathway as a therapeutic target.

The paper identifies an epigenetic neural signature in glioblastoma that correlates with poor prognosis and therapeutic vulnerability.

The paper compares TCR-pMHC prediction models and finds they lack generalization and robustness due to data dependency.

Pathonoia detects viruses and bacteria in RNA-seq data with high precision and recall, outperforming existing methods and providing insights into microbe-host interactions.

This study uses single-cell RNA sequencing to identify a mechanosensitive MRTF-SRF pathway in early diabetic nephropathy, validated in human datasets and ex vivo experiments.

The paper identifies Err2 and Err3 as transcription factors promoting functional properties in gamma motor neurons essential for proprioceptive movement control.

The paper identifies early translatome changes in specific neural cell types before EEG and neuropathological signs of prion disease in mice.

The paper shows that neuronal cholesterol synthesis is critical for remyelination in chronic demyelination models and MS patients.

This paper introduces a deep-learning based domain adaptation algorithm for RNA-seq metadata annotation, achieving higher accuracy than existing methods for tissue, sample source, and patient sex prediction.

Scaden is a deep learning model for cell deconvolution that outperforms existing methods and works across bulk RNA-seq, microarray, and human/mouse tissue data.

This paper introduces cscGANs for generating realistic single cell RNA-seq data to augment sparse datasets, improving analysis outcomes and reducing experimental costs.

SEA provides a web tool for querying and analyzing small RNA expression across ten organisms, including differential expression and classification comparisons.

This study uses genome-wide association data for educational attainment and schizophrenia to identify genetic heterogeneity and novel candidate genes for schizophrenia.

Oasis2.0 improves small RNA-seq analysis by enhancing detection, classification, and visualization of small RNAs across organisms and identifying cross-species miRNAs and viral/bacterial sRNAs.

The paper estimates the extent and biological role of mutually exclusive splicing in humans by identifying and validating over 855 MXEs, including novel exons, and links them to disease pathology.