MacroGuide: Topological Guidance for Macrocycle Generation

TL;DR

MacroGuide introduces a topological guidance method using Persistent Homology to significantly improve macrocycle generation rates in molecular generative models, achieving near-perfect macrocycle production while maintaining quality.

Contribution

It presents a novel diffusion guidance mechanism leveraging topological features to enhance macrocycle generation in deep generative models.

Findings

MacroGuide increases macrocycle generation from 1% to 99%.

The method matches or exceeds state-of-the-art quality metrics.

It effectively guides models in both unconditional and conditional settings.

Abstract

Macrocycles are ring-shaped molecules that offer a promising alternative to small-molecule drugs due to their enhanced selectivity and binding affinity against difficult targets. Despite their chemical value, they remain underexplored in generative modeling, likely owing to their scarcity in public datasets and the challenges of enforcing topological constraints in standard deep generative models. We introduce MacroGuide: Topological Guidance for Macrocycle Generation, a diffusion guidance mechanism that uses Persistent Homology to steer the sampling of pretrained molecular generative models toward the generation of macrocycles, in both unconditional and conditional (protein pocket) settings. At each denoising step, MacroGuide constructs a Vietoris-Rips complex from atomic positions and promotes ring formation by optimizing persistent homology features. Empirically, applying MacroGuide to pretrained diffusion models increases macrocycle generation rates from 1% to 99%, while matching or exceeding state-of-the-art performance on key quality metrics such as chemical validity, diversity, and PoseBusters checks.