ConforNets: Latents-Based Conformational Control in OpenFold3

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lee, Minji, Kalicki, Colin, Jeon, Minkyu, Qabel, Aymen, Fadini, Alisia, AlQuraishi, Mohammed
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913048318443520
author Lee, Minji
Kalicki, Colin
Jeon, Minkyu
Qabel, Aymen
Fadini, Alisia
AlQuraishi, Mohammed
author_facet Lee, Minji
Kalicki, Colin
Jeon, Minkyu
Qabel, Aymen
Fadini, Alisia
AlQuraishi, Mohammed
contents Models from the AlphaFold (AF) family reliably predict one dominant conformation for most well-ordered proteins but struggle to capture biologically relevant alternate states. Several efforts have focused on eliciting greater conformational variability through ad hoc inference-time perturbations of AF models or their inputs. Despite their progress, these approaches remain inefficient and fail to consistently recover major conformational modes. Here, we investigate both the optimal location and manner-of-operation for perturbing latent representations in the AF3 architecture. We distill our findings in ConforNets: channel-wise affine transforms of the pre-Pairformer pair latents. Unlike previous methods, ConforNets globally modulate AF3 representations, making them reusable across proteins. On unsupervised generation of alternate states, ConforNets achieve state-of-the-art success rates on all existing multi-state benchmarks. On the novel supervised task of conformational transfer, ConforNets trained on one source protein can induce a conserved conformational change across a protein family. Collectively, these results introduce a mechanism for conformational control in AF3-based models.
format Preprint
id arxiv_https___arxiv_org_abs_2604_18559
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle ConforNets: Latents-Based Conformational Control in OpenFold3
Lee, Minji
Kalicki, Colin
Jeon, Minkyu
Qabel, Aymen
Fadini, Alisia
AlQuraishi, Mohammed
Biomolecules
Machine Learning
Models from the AlphaFold (AF) family reliably predict one dominant conformation for most well-ordered proteins but struggle to capture biologically relevant alternate states. Several efforts have focused on eliciting greater conformational variability through ad hoc inference-time perturbations of AF models or their inputs. Despite their progress, these approaches remain inefficient and fail to consistently recover major conformational modes. Here, we investigate both the optimal location and manner-of-operation for perturbing latent representations in the AF3 architecture. We distill our findings in ConforNets: channel-wise affine transforms of the pre-Pairformer pair latents. Unlike previous methods, ConforNets globally modulate AF3 representations, making them reusable across proteins. On unsupervised generation of alternate states, ConforNets achieve state-of-the-art success rates on all existing multi-state benchmarks. On the novel supervised task of conformational transfer, ConforNets trained on one source protein can induce a conserved conformational change across a protein family. Collectively, these results introduce a mechanism for conformational control in AF3-based models.
title ConforNets: Latents-Based Conformational Control in OpenFold3
topic Biomolecules
Machine Learning
url https://arxiv.org/abs/2604.18559