José M. Duarte

This paper introduces a scalable method for searching 3D protein structure similarity across entire proteomes using deep learning and vector databases.

The paper introduces a method for real-time retrieval of protein structures from the PDB by comparing electron density volumes using 3D Zernike moments.

The paper presents a method to assess the quaternary structure of proteins from crystal data by combining interface classification with symmetry and topology analysis, achieving 85% precision on a new dataset.

The paper evaluates quaternary structure annotations in the PDB by comparing prediction methods and develops a web tool for researchers.

The paper introduces MMTF, a new binary format for macromolecular structures, which is faster and smaller than PDBx/mmCIF, enabling efficient visualization and analysis.