Cameron Mura

The paper introduces DeepUrfold, a deep generative model of protein structure that reveals continuous relationships in fold space beyond traditional discrete classifications.

Prop3D provides a platform for creating and sharing featurized protein domain datasets, with a focus on avoiding data leakage through evolutionary relationships.

The paper introduces a framework combining deep generative models and layerwise relevance propagation to identify ancestral peptide fragments in protein structures, revealing evolutionary relationships.

PortalCG predicts chemical-protein interactions for dark proteins using a deep learning framework that integrates binding site information and meta-learning, outperforming existing methods in out-of-distribution scenarios.

The paper introduces the Urfold concept to explore a new level of protein structural granularity between architecture and topology, aiming to reconcile structural similarity and fold space continuity.

The paper provides a survey-based analysis of the structural properties and functional diversity of small β-barrel domains.