MAGNet: Motif-Agnostic Generation of Molecules from Shapes
This work addresses the problem of inflexibility in molecule generation for drug discovery by enabling representation of substructures beyond known motifs, though it appears incremental as it builds on existing graph-based methods.
The paper tackled the limitation of molecule generation models that rely on predefined motifs by proposing MAGNet, a graph-based model that generates abstract shapes before assigning atom and bond types, resulting in outperforming most other graph-based approaches on standard benchmarks and producing molecules with more topologically distinct structures and diverse assignments.
Recent advances in machine learning for molecules exhibit great potential for facilitating drug discovery from in silico predictions. Most models for molecule generation rely on the decomposition of molecules into frequently occurring substructures (motifs), from which they generate novel compounds. While motif representations greatly aid in learning molecular distributions, such methods struggle to represent substructures beyond their known motif set. To alleviate this issue and increase flexibility across datasets, we propose MAGNet, a graph-based model that generates abstract shapes before allocating atom and bond types. To this end, we introduce a novel factorisation of the molecules' data distribution that accounts for the molecules' global context and facilitates learning adequate assignments of atoms and bonds onto shapes. Despite the added complexity of shape abstractions, MAGNet outperforms most other graph-based approaches on standard benchmarks. Importantly, we demonstrate that MAGNet's improved expressivity leads to molecules with more topologically distinct structures and, at the same time, diverse atom and bond assignments.