Inverse FoldDir: Structure-conditioned Protein Sequence Design by Dirichlet Flow Matching
- 1Stanford University
Abstract
Protein engineering has important implications in the bioeconomy, enabling applications in materials, medicine, and energy. A key challenge is designing protein sequences that have a specific form and function. Protein inverse folding seeks to address this challenge by identifying amino acid sequences compatible with a desired protein backbone. This task is central to protein redesign and can provide a sequence design capability for de novo backbones produced by structure-generation methods. Ideally, inverse folding can provide diverse sequence alternatives, fixed residues or motifs, soft biochemical preferences at selected positions, and candidates that remain experimentally useful. We developed Inverse FoldDir, a controllable inverse-folding method that performs iterative denoising on the amino acid probability simplex. Given a backbone structure, the model updates all positions jointly through a learned Dirichlet flow, supporting full sequence generation, fixed-residue inpainting, and user-defined soft residue priors. On the held-out CATH 4.2 test set, Inverse FoldDir achieved a mean TM-score of 84.5 (on a 0-100 scale) and a mean C RMSD of 1.76[A], compared with 83.3 and 1.86[A], respectively, for ESM-IF1, the strongest evaluated baseline on both metrics. Denoising trajectory analyses showed that positions commit at different rates and that some residues change identity late in generation, illustrating whole-sequence refinement rather than one-shot prediction or irreversible sequential decoding. We experimentally tested Inverse FoldDir in an anti-GFP nanobody redesign task, where two of 35 redesigned sequences retained reproducible sfGFP-binding signal across independent assay runs with approximately 43% sequence divergence from the native nanobody. Inverse FoldDir is a structure-conditioned protein redesign method that combines structural recovery, user control, experimental validation, and a natural route toward future property-guided sampling.
Related
- RiboDFM: Robust RNA sequence design using backbone-conditioned discrete flow matching
- TopoFlow: Evolutionarily Conditioned Flow Matching for Protein Conformational Ensemble Generation
- TempoBridge: Source-Conditioned Flow Matching with Optimal Transport Couplings for Single-Cell Population Transitions
- GeoSinkhorn Flow: Geometry-Aware Flow Matching for Perturbation-Conditioned Dynamics in Single cell Data Phenoscapes
- MerKurio: sequence extraction and annotation based on matching k-mers
BibTeX
@misc{tartici2026inverse,
title = {{Inverse FoldDir: Structure-conditioned Protein Sequence Design by Dirichlet Flow Matching}},
author = {A. TARTICI and M. Stojkovic and A. Tian and M. C. Jewett and R. B. Altman and B. J. Wittmann},
howpublished = {bioRxiv},
year = {2026},
month = sep,
doi = {10.64898/2026.09.06.749733},
url = {https://www.biorxiv.org/content/10.64898/2026.09.06.749733v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Inverse FoldDir: Structure-conditioned Protein Sequence Design by Dirichlet Flow Matching
AU - A. TARTICI
AU - M. Stojkovic
AU - A. Tian
AU - M. C. Jewett
AU - R. B. Altman
AU - B. J. Wittmann
PY - 2026
DA - 2026/09/06
PB - bioRxiv
DO - 10.64898/2026.09.06.749733
UR - https://www.biorxiv.org/content/10.64898/2026.09.06.749733v1
AB - Protein engineering has important implications in the bioeconomy, enabling applications in materials, medicine, and energy. A key challenge is designing protein sequences that have a specific form and function. Protein inverse folding seeks to address this challenge by identifying amino acid sequences compatible with a desired protein backbone. This task is central to protein redesign and can provide a sequence design capability for de novo backbones produced by structure-generation methods. Ideally, inverse folding can provide diverse sequence alternatives, fixed residues or motifs, soft biochemical preferences at selected positions, and candidates that remain experimentally useful. We developed Inverse FoldDir, a controllable inverse-folding method that performs iterative denoising on the amino acid probability simplex. Given a backbone structure, the model updates all positions jointly through a learned Dirichlet flow, supporting full sequence generation, fixed-residue inpainting, and user-defined soft residue priors. On the held-out CATH 4.2 test set, Inverse FoldDir achieved a mean TM-score of 84.5 (on a 0-100 scale) and a mean C RMSD of 1.76[A], compared with 83.3 and 1.86[A], respectively, for ESM-IF1, the strongest evaluated baseline on both metrics. Denoising trajectory analyses showed that positions commit at different rates and that some residues change identity late in generation, illustrating whole-sequence refinement rather than one-shot prediction or irreversible sequential decoding. We experimentally tested Inverse FoldDir in an anti-GFP nanobody redesign task, where two of 35 redesigned sequences retained reproducible sfGFP-binding signal across independent assay runs with approximately 43% sequence divergence from the native nanobody. Inverse FoldDir is a structure-conditioned protein redesign method that combines structural recovery, user control, experimental validation, and a natural route toward future property-guided sampling.
ER -