Klaus Liedl

University of Innsbruck, Austria - Research

From Canonical Structures to Conformational Ensembles: Physics-Based Modeling of Antibody Structure, Dynamics, and Developability in Solution

Abstract

Therapeutic antibodies constitute one of the fastest growing classes of biologics, yet their function is governed by conformational dynamics that are poorly captured by static crystal structures or canonical loop classifications. Over the past three decades, antibody modeling has evolved from knowledge-based approaches toward physics-based molecular simulations that explicitly describe motions in solution. Here, we present a computational framework that combines atomistic molecular dynamics, enhanced sampling, and kinetic state models to resolve conformational ensembles of antibody complementarity-determining regions (CDRs) on micro- to millisecond timescales. These ensembles replace single “canonical” structures and enable quantitative predictions of binding-competent states, paratope geometries, and mechanisms ranging from conformational selection to induced fit. Incorporating improved force fields, polarization effects, and rigorous solvation thermodynamics allows robust free-energy estimates and reliable structure–property relationships. We demonstrate how ensemble-based descriptions improve docking accuracy, rationalize humanization and framework effects, and quantify developability features such as electrostatics and hydrophobicity using grid inhomogeneous solvation theory. Across diverse systems, physics-based predictions of loop conformations and biophysical properties are reproducible and computationally tractable, with typical turnaround times of one to two weeks per antibody. Together, these advances establish conformational ensembles in solution as a new paradigm for antibody modeling, bridging molecular physics with therapeutic design and enabling predictive, mechanistic insights for next-generation biologics.