Cheminformatics

Generative Molecular Morphing for Flexible-Size Design via Unbalanced Optimal Transport

Malte Franke, Stefan P. Schmid, Zarko Ivkovic, Kjell Jorner, Andreas Krause

Featured June 10, 2026

This analysis was generated by SciGrove. Upload your own PDFs or enter a DOI — and get the same AI breakdown on any paper.

Get started

AI-generated analysis — This is SciGrove's AI interpretation of the paper, not peer-reviewed content. Always refer to the original paper.

Simply

A new AI model called Morph can design molecules by adding or removing atoms, unlike older methods that only changed fixed-size molecules, allowing it to create completely new chemical structures.

In depth

The paper introduces Morph, a novel generative model that overcomes the fixed-size limitation of previous 3D molecular design methods. It achieves this by formulating molecular generation as a jump-flow process on geometric graphs, allowing for dynamic insertions, deletions, and substitutions of atoms. A key innovation is the use of unbalanced optimal transport to align molecules of different sizes, enabling the definition of smooth interpolation paths for training the generative process.

Key Takeaways

1
The `Morph` model enables flexible-size 3D molecular design by dynamically adapting the number of atoms during generation, a significant advancement over fixed-size models.
2
It leverages unbalanced optimal transport (UOT) to establish structural alignments between molecules of varying sizes, which is crucial for defining interpolation paths for training.
3
The model demonstrates superior property steerability and the ability to generate valid molecules far out-of-distribution (OOD), including scaffold decoration, which is vital for accelerating chemical discovery.

Conceptual Flow

HIGH LEVEL

Methodology: How Morph Designs Molecules

The model learns to change simple starting molecules into complex new ones by adding, removing, or swapping parts, like building with LEGOs.

Simple Start Molecule

Add / Remove / Change Parts

Complex New Molecule

Results: Creating New Chemical Designs

This new method can make molecules of different sizes and with specific desired traits, even ones it hasn't seen before, helping scientists find new drugs faster.

Desired Properties

Generate Molecules

New Molecules

Custom Sizes

Unique Features