[Abstract - QISK 2026] DockMaster-Songdo™: A Three-Level Quantum Adoption Framework for Protein Preprocessing

DockMaster-Songdo A Three-Level Quantum Adoption Framework for Protein Preprocessing_2026 Anuual meeting QISK.pdf

Download PDF • 244KB

This paper presents a three-level quantum adoption framework for upgrading DockMaster-Songdo™, a web-based protein preprocessing platform.

Current State (Classical/AI Pipeline)

DockMaster-Songdo™ currently automates six protein structure preprocessing steps using rule-based and LLM-driven workflows:

• Water removal

• Hydrogen addition

• Charge assignment

• Ligand and metal handling

• PDBQT export

Proposed Framework

The framework transforms the existing classical/AI pipeline into three levels:

Level 0: Pure classical baseline

Level 1: Bio-physics-inspired enhancement

Level 2: Quantum-optional hybrid

Quantum Algorithm Mapping

Each preprocessing step is mapped to specific near-term quantum algorithms:

• QAOA-assisted water microstate free-energy estimation

• Fragment-VQE for protonation microstates

• FMO/VQE-based charge assignment

• Quantum-kernel SVM for ligand relevance classification

• Metal-center VQE for metal handling

• VQE-assisted torsional degree-of-freedom annotation for PDBQT export

Design Approach

The paper emphasizes:

• Algorithm selection based on biophysical interpretability, qubit/depth scaling, and NISQ (Noisy Intermediate-Scale Quantum) feasibility

• Exclusion of QUBO/annealing-based formulations

• Preference for fragment-scale electronic-structure and quantum-ML primitives

#QuantumComputing #AI #LLM #Protein #NewDrug #Design #Development