pCoMole: Pareto-Constrained Molecule Editing with Discrete Flows

2026-10-01

AI Summary

This study addresses the joint challenges of multi-objective optimization, hard constraint satisfaction, and discrete variable-length sequence generation in biomolecular editing. To this end, it proposes a guided generation framework based on discrete flow matching that modulates pretrained models via the Doob h-transform. Furthermore, by designing a feasibility-gated terminal distribution combined with a Tchebycheff utility function and Monte Carlo approximation, the method synergistically ensures Pareto preference alignment and biochemical constraint satisfaction. The proposed approach successfully achieves green fluorescent protein (GFP) compression while preserving fluorescence, as well as Cas9 shortening while maintaining specificity. Wet-lab experiments further validate the functional effectiveness of the designed sequences.

Abstract

Biomolecular therapeutics often start from known sequences and require targeted editing to improve multiple properties while satisfying hard biochemical and manufacturability constraints. However, existing generative methods do not jointly support multi-objective optimization, hard feasibility, and sequence editing in discrete, variable-length biological spaces. In this work, we introduce Pareto-Constrained Molecule Editing (pCoMole), a framework built on discrete flow matching that steers a pre-trained Edit Flow toward user-specified preferences while enforcing terminal feasibility. pCoMole defines a feasibility-gated terminal distribution using an augmented Tchebycheff utility and realizes the resulting preference tilt through a Doob-h transform of the underlying edit process. To make this construction practical, we approximate the required harmonic function using short Monte Carlo rollouts over candidate edits, yielding an efficient guided editor with provable preference consistency. We validate pCoMole by shrinking GFP while retaining fluorescence-related properties, shortening diverse Cas9 orthologs while preserving PAM specificity, and compressing peptide binders into short peptidomimetics that optimize seven drug-related properties under hard constraints. In wet lab testing, two 229-residue pCoMole-designed eGFP variants retained clear green fluorescence in BL21 cells after 10 deletions, with either one or two substitutions. Together, pCoMole enables constraint-aware, Pareto-aligned editing of biomolecular sequences in discrete, variable-length spaces.

Problem

• molecule editing
• multi-objective optimization
• hard constraints
• discrete sequence space
• biomolecular therapeutics

Innovation

• Discrete Flow Matching
• Pareto-Constrained Optimization
• Doob-h Transform
• Biomolecule Editing
• Monte Carlo Rollouts