Peptide modeling and biomolecular recognition

The emerging Pep direction extends my work from reaction informatics and molecular machine learning toward biomolecular interactions, confidence-aware scoring, peptide–protein complex analysis, and computational peptide design.

Key themes

• peptide–protein complex querying and post-processing

• interaction analysis and confidence-aware scoring

• structure-guided peptide design workflows

• reusable computational infrastructure for peptide science

This direction is motivated by a broader scientific question: how can we connect structure, interaction geometry, confidence, and biological relevance in computational workflows for peptide discovery? Current tools in this direction include PepKit and PepQ, as well as collaborative work on cyclic peptide binders and target-focused peptide modeling projects.

Why it matters

Peptides sit at an important interface between chemistry and molecular biology. Better computational workflows for peptide recognition, ranking, and refinement can support both therapeutic discovery and mechanistic chemical biology.