[Research] Multi-Platform Quantum Computing Simulation Pipeline for ADC Drug Design: HER2-Trastuzumab Interface Analysis

Multi-Platform Quantum Computing Simulation Pipeline for ADC Drug Design HER2-Trastuzumab Interface Analysis_20260406.pdf

Download PDF • 980KB

Photonic Screening, Neutral-Atom Optimization, and Gate-Based Quantum Chemistry for Antibody-Drug Conjugate Binding Site Identification

Research Report | April 2026 | Author: Sungil Oh | os81paul@gmail.com

Korean Society of Medical & Biological Engineering (Member) | Quantum Information Society of Korea (Member)

Keywords: Quantum Computing Simulation, Antibody-Drug Conjugate (ADC), HER2-Trastuzumab, Boson Sampling, MIS Rydberg Blockade, BEAST-VQE, Projective Embedding, PDB 1N8Z, Quandela, Pasqal, Kvantify Qrunch

This report extends the preliminary work presented at the 2026 Spring Conference of the Korean Society of Medical & Biological Engineering ("Photonic Boson-Sampling Simulation-Based Interface Analysis of the HER2-Trastuzumab Complex for ADC Research: A Feasibility Study", Sungil Oh., 2026), which demonstrated Stage 1 (Quandela, 12-mode, 22 residues). The present work expands to a full 3-stage pipeline with 49 residues, three quantum platforms, PDB-derived binding energy calculations, and an interactive web application.

Executive Summary

This report presents a novel three-stage quantum computing pipeline for antibody-drug conjugate (ADC) drug design, applied to the HER2-Trastuzumab interface (PDB 1N8Z). The pipeline integrates three distinct quantum computing paradigms: photonic (Quandela Perceval), neutral-atom (Pasqal Pulser), and gate-based (Kvantify Qrunch) quantum platforms, each contributing a specialized computational capability to the drug design workflow.

Key Results: Stage 1 (Boson Sampling) screened all 49 interface residues in 1.87 seconds, identifying binding hotspots with high contrast. Stage 2 (MIS Rydberg Blockade) selected 16 structurally independent binding sites via 360-degree structural optimization. Stage 3 (BEAST-VQE) calculated a covalent binding energy of -25.6 kcal/mol using PDB-derived coordinates with B3LYP/6-31G* and Projective Embedding, confirming thermodynamic favorability. The identified binding sites (PRO571, GLY103, PRO572, ASP570) match experimentally confirmed drug sites from Cho et al. (2003, Nature).

All simulations were performed on classical simulators (state-vector), which produce results mathematically identical to quantum processing units (QPUs) at the 10-qubit scale used in this study. The complete pipeline is deployed as an interactive web application with 3D molecular visualization, animated quantum computation processes, and publication-quality comparison charts.