Towards Conditional Generation of Minimal Action Potential Pathways for\n Molecular Dynamics

2021/11/28 by John Kevin Cava, Cava, John Kevin, John Vant +11 · 1 citation
Biochemistry, Genetics and Molecular Biology · Computer Science · Physics and Astronomy · #Algorithms and Data Compression #Artificial Intelligence (cs.AI) #Biological Physics (physics.bio-ph) #Biomolecules (q-bio.BM) #FOS: Biological sciences #FOS: Computer and information sciences #FOS: Physical sciences #Machine Learning (cs.LG) #Protein Structure and Dynamics #Scientific Research and Discoveries

paper · pdf · doi:10.48550/arxiv.2111.14053

openalex publication_date 2021/11/28 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

In this paper, we utilized generative models, and reformulate it for problems\nin molecular dynamics (MD) simulation, by introducing an MD potential energy\ncomponent to our generative model. By incorporating potential energy as\ncalculated from TorchMD into a conditional generative framework, we attempt to\nconstruct a low-potential energy route of transformation between the\nhelix~\→~coil structures of a protein. We show how to add an\nadditional loss function to conditional generative models, motivated by\npotential energy of molecular configurations, and also present an optimization\ntechnique for such an augmented loss function. Our results show the benefit of\nthis additional loss term on synthesizing realistic molecular trajectories.\n

Citations

• Anisotropy of Fluctuation Dynamics of Proteins with an Elastic Network Model
• Conditional Generative Adversarial Nets

Cited by

• Embedding Trust at Scale: Physics-Aware Neural Watermarking for Secure and Verifiable Data Pipelines

Related

• Follow the MEP: Scalable Neural Representations for Minimum-Energy Path Discovery in Molecular Systems
• Programmable Potentials: Approximate N-body potentials from coarse-level\n logic
• Generative Modeling of Molecular Dynamics Trajectories
• Minima hopping: Searching for the global minimum of the potential energy surface of complex molecular systems without invoking thermodynamics
• Molecular dynamics without molecules: searching the conformational space of proteins with generative neural networks