3D Molecular Generation by Virtual Dynamics

Shuqi Lu, Lin Yao, Xi Chen, Hang Zheng, Guolin Ke

Keywords: 3D molecular generation, structure-based drug design

Abstract: Structure-based drug design, i.e., finding molecules with high affinities to the target protein pocket, is one of the most critical tasks in drug discovery. Traditional solutions, like virtual screening, require exhaustively searching on a large molecular database, which are inefficient and cannot get novel molecules beyond the database. The pocket-based 3D molecular generation model, i.e., directly generating a molecule with a 3D structure and binding position in the pocket, is a new promising way to address this issue. However, the method is very challenging due to the complexity brought by the huge continuous 3D space in the pocket cavity. Herein, inspired by Molecular Dynamics, we propose a novel pocket-based 3D molecular generation framework VD-Gen. VD-Gen consists of a Virtual Dynamics mechanism and several carefully designed stages to generate fine-grained 3D molecules with binding positions in the pocket cavity end-to-end. Rather than directly generating or sampling atoms with 3D positions in the pocket like in early attempts, in VD-Gen, we first randomly scatter many virtual particles in the pocket; then with the proposed Virtual Dynamics mechanism, a deep model, acting like a "force field", iteratively moves these virtual particles to positions that are highly possible to contain real atoms. After virtual particles are stabilized in 3D space, we extract the atoms from them. Finally, we further refine the 3D positions of atoms by Virtual Dynamics again, to get a fine-grained 3D molecule. Extensive experiment results on pocket-based molecular generation demonstrate that VD-Gen can generate novel 3D molecules to fill the target pocket cavity with high binding affinities, significantly outperforming previous baselines.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Machine Learning for Sciences (eg biology, physics, health sciences, social sciences, climate/sustainability )

TL;DR: We propose a novel pocket-based 3D molecular generation framework VD-Gen, which consists of a Virtual Dynamics mechanism and several carefully designed stages to generate fine-grained molecules with binding positions in the pocket cavity end-to-end.