Jens Meiler

This paper introduces a differentiable spin-center diffusion model to generate weighted protein ensembles from DEER spectroscopy data, enabling thermodynamic interpretation of conformational ensembles.

AltereGO fuses protein language models and structure representations to predict molecular functions, achieving better performance than individual models and some existing methods.

The paper validates and analyzes 12,000 AI-designed CAR-T binders from a global competition, identifying successful designs and factors improving their performance.

This paper introduces RosettaLigand PathFinder, a tool for simulating ligand trajectories through proteins, showing it can reproduce known transport features and mutation effects.

This study evaluates VX-121's efficacy across 232 rare CFTR variants, finding it generally more effective than VX-445 but with one variant showing reduced selectivity.

The paper develops a compound prioritization strategy using a multi-layer perceptron trained on docked poses from multiple docking approaches to identify putative binders for the nuclear receptor LRH-1.

VUStruct analyzes genetic variants of unknown significance by integrating 3D protein structure data and machine learning to improve clinical and research interpretation of rare genetic disorders.

This study identifies key Y4R residues and discovers selective non-peptide small-molecule agonists through molecular docking, residue replacement, and ultra-large library screening.

The paper determines the structure of the Gβγ-prefusion SNARE complex using cryo-EM and proposes a model for how Gβγ inhibits synaptic vesicle fusion by blocking SNARE complex assembly.

FrustraMPNN rapidly predicts single-residue frustration in proteins at proteome scale with high accuracy, enabling new insights into protein function and dynamics.

The paper designs and experimentally validates mini-protein binders targeting receptor-binding domains of Nipah, Langya, and Measles viruses using computational tools and limited screening.

The paper develops classifiers to distinguish functional and trafficking effects of KCNQ1 variants, improving variant interpretation for LQTS and channelopathies.

This study identifies a small molecule, VU0494372, that enhances KCNQ1 expression and trafficking, offering a potential therapeutic approach for LQTS.

SuperMetal predicts metal-binding sites in proteins with high precision and efficiency, outperforming existing models for zinc ions.

This paper evaluates combining graph neural networks with expert descriptors for ligand-based virtual screening, finding that simpler GNNs with descriptors perform comparably to complex ones and that expert descriptors excel in scaffold-split scenarios.

The paper uses proteomics and computational modeling to analyze CFTR variants, identifying structural vulnerabilities in NBD1 that limit corrector efficacy and suggesting secondary mutations as a therapeutic strategy.

This study investigates how calnexin modulates the expression and pharmacological rescue of CFTR variants using deep mutational scanning, revealing its role in CFTR assembly and variant-specific corrector responses.

FakeRotLib efficiently parameterizes non-canonical amino acids in Rosetta by generating rotamer distributions from small molecule conformers, outperforming existing methods.

The paper identifies 57 druggable targets for MASLD using genetic data and evaluates their therapeutic potential through Mendelian randomization and structural modeling.

This study compares self-supervised ML methods with biophysical approaches in protein design, finding ML improves sampling but not scoring of high-fitness variants.

This study designs improved VEGFA inhibitors using computational tools, achieving higher affinity than existing peptides with minimal experimental rounds.

This paper introduces HyperMPNN, a deep learning model trained on hyperthermophile proteins to design highly thermostable proteins, achieving stability up to 95°C.

The paper develops a machine learning model to predict MHC-I binding affinity for peptides with non-canonical amino acids and evaluates its performance against existing methods.

SuperWater predicts water molecule positions on protein structures using a generative AI framework, achieving high accuracy in crystal water coverage and prediction precision.

The paper evaluates computational hit-finding methods for targeting the WDR domain of LRRK2, a Parkinson's disease protein, finding that diverse strategies yielded weakly potent compounds.

MutationExplorer provides a webserver to visualize the energetic impacts of protein mutations on 3D structures using Rosetta and supports SNP analysis and sequence uploads.

This study benchmarks AlphaMissense's ability to predict pathogenicity of CFTR missense variants using clinical and functional metrics, finding moderate correlations with some outcomes but limited utility for pharmacological response.

The paper introduces a chirality-aware graph neural network for drug discovery that improves predictive performance and interpretability in quantitative structure activity relationship modeling.

The paper uses personalized structural biology to analyze de novo KCNC2 variants in a DEE patient, revealing distinct functional effects of two nearby variants.