MolAlign3D: Enhancing Fixed-Dimensional E(3)-Equivariant Latent Space for High-Fidelity 3D Molecular Reconstruction and Editing

Recent advances in 3D molecular modeling have achieved high-fidelity structural synthesis, yet these models often lack an explicit and manipulable representation space. To address this, MolFLAE introduced a fixed-dimensional, E(3)-equivariant latent space, providing a novel framework for molecular editing independent of atom counts. However, because its latent space was primarily optimized for geometric reconstruction, it remains semantically shallow and inadequate for comprehensive representation learning. In this work, we propose MolAlign3D, which evolves this architecture into a unified semantic-generative engine. By anchoring MolFLAE’s manipulable latents with embeddings from a pre-trained molecular encoder, we yield a manifold that is both semantically dense and geometrically precise. Experiments show that MolAlign3D achieves high-fidelity molecular reconstruction and attains comparable performance on molecular property prediction benchmarks. Notably, the integration of rich semantic priors significantly enhances zero-shot molecular manipulation, including atom-number editing and latent-space interpolation, outperforming prior fixed-dimensional equivariant latent baseline.

AI can now build accurate 3D molecular structures, but these models are black boxes. Researchers cannot easily see inside or control how the model works. We developed MolAlign3D, which links a shape focused inner space to a pretrained system that understands chemical meaning. This gives the model both spatial accuracy and chemical insight, allowing it to grasp what molecules mean, not just how they look. This combined understanding enables precise editing of molecules on the fly, such as adding or removing atoms and smoothly transforming one molecule into another, beating previous methods for guided design in drug discovery.