Ensembles

Jul 28, 2026

Protein backbone design diffusion models can be repurposed for fitting structures into electron density

Apr 21, 2026

Diffusion-based structure prediction can be steered by modifying the conditioning embeddings rather than the latent space, and such embeddings can be used for subsequent iterations

Apr 21, 2026

AF2 RMSD to ground truth correlated with RMSD between conformations

Apr 21, 2026

AlphaFlow and ESMFlow

Apr 21, 2026

AlphaFold2 and RosettaFold sometimes produce different conformations

prediction/ensembles

Apr 21, 2026

AlphaFold2 predicts Apo conformations with poorer RMSD than holo conformations

prediction/ensembles

Apr 21, 2026

AlphaFold2 predicts ground states more often

prediction/ensembles

Apr 21, 2026

AlphaFold2 predicts the holo form of proteins in most cases

prediction/ensembles

Apr 21, 2026

AlphaFold3 performs comparably to AlphaFold2 when predicting multiple conformations of fold-switching proteins

prediction/ensembles

Apr 21, 2026

AlphaFold3 universally predicts the active state of ligand-bound GPCRs, even when the ligand is an antagonist

Apr 21, 2026

Alternate conformations can be sampled by inverting AlphaFold2 and backpropagating along specific user-picked collective variables

Apr 21, 2026

Alternate conformations can be sampled with MSA-based structure prediction methods using custom PDB databases and subsampled MSAs

Apr 21, 2026

Conformationally stabilizing mutations can be identified from sequence clusters that are predicted to fold into that conformation when passed to AlphaFold2

Apr 21, 2026

Diffusion-based structure prediction can sometimes model conformational ensembles

Apr 21, 2026

Enhanced diffusion with metadynamics-like potentials can sometimes convergence slower than unbiased diffusion

Apr 21, 2026

Ensembles can be modeled by structure prediction NNs using experimental data via guidance sampling

Apr 21, 2026

Inverse folding NNs are better predictors of equilibrium dynamics than protein folding NNs

Apr 21, 2026

Protein ensemble prediction methods do not generate conformations that are amenable to PPI docking

Apr 21, 2026

Protein folding neural networks can be combined with VAEs for conformational modeling

Apr 21, 2026

Protein structure prediction methods are unable to predict the energetics of a conformational landscape unless explicitly trained for that purpose

Apr 21, 2026

Structure ensemble prediction methods are unable to accurately model subtle conformational differences captured by X-ray crystallography

Apr 21, 2026

Structure prediction from conformational samplers tends to be worse than those from vanilla structure prediction methods

prediction/ensembles

Apr 21, 2026

Structure prediction methods undersample the conformational space they find to be high-confidence

Apr 21, 2026

Structure prediction neural networks predict different conformations of a given sequence depending on if a single chain is predicted versus multiple chains

Apr 21, 2026

Structure prediction neural networks trained on the same structural dataset are likely to predict the same conformation

Apr 21, 2026

Structure prediction with AlphaFold improved when starting from an initial guess, rather than default initialization

Apr 21, 2026

The AlphaFold and ESMFold databases can be mined for alternate conformations

Apr 21, 2026

pTM is less lenient than pLDDT toward structural novelty when predicting alternate conformations

Apr 21, 2026

Biomolecular diffusion models cannot reproduce the equilibrium dynamics of the simulations they are trained on

Apr 21, 2026

Conformational sampling by AlphaFold3-generation methods can be achieved by scaling the pair representation

Apr 21, 2026

Deep learning methods produce different CDRH3 conformers

Apr 21, 2026

Guidance potentials can be added to diffusion-based structure prediction for enhanced sampling of protein conformations

Apr 21, 2026

Quartz 4

Explorer

Ensembles

Protein backbone design diffusion models can be repurposed for fitting structures into electron density

Diffusion-based structure prediction can be steered by modifying the conditioning embeddings rather than the latent space, and such embeddings can be used for subsequent iterations

AF2 RMSD to ground truth correlated with RMSD between conformations

AlphaFlow and ESMFlow

AlphaFold2 and RosettaFold sometimes produce different conformations

AlphaFold2 predicts Apo conformations with poorer RMSD than holo conformations

AlphaFold2 predicts ground states more often

AlphaFold2 predicts the holo form of proteins in most cases

AlphaFold3 performs comparably to AlphaFold2 when predicting multiple conformations of fold-switching proteins

AlphaFold3 universally predicts the active state of ligand-bound GPCRs, even when the ligand is an antagonist

Alternate conformations can be sampled by inverting AlphaFold2 and backpropagating along specific user-picked collective variables

Alternate conformations can be sampled with MSA-based structure prediction methods using custom PDB databases and subsampled MSAs

Conformationally stabilizing mutations can be identified from sequence clusters that are predicted to fold into that conformation when passed to AlphaFold2

Diffusion-based structure prediction can sometimes model conformational ensembles

Enhanced diffusion with metadynamics-like potentials can sometimes convergence slower than unbiased diffusion

Ensembles can be modeled by structure prediction NNs using experimental data via guidance sampling

Inverse folding NNs are better predictors of equilibrium dynamics than protein folding NNs

Protein ensemble prediction methods do not generate conformations that are amenable to PPI docking

Protein folding neural networks can be combined with VAEs for conformational modeling

Protein structure prediction methods are unable to predict the energetics of a conformational landscape unless explicitly trained for that purpose

Structure ensemble prediction methods are unable to accurately model subtle conformational differences captured by X-ray crystallography

Structure prediction from conformational samplers tends to be worse than those from vanilla structure prediction methods

Structure prediction methods undersample the conformational space they find to be high-confidence

Structure prediction neural networks predict different conformations of a given sequence depending on if a single chain is predicted versus multiple chains

Structure prediction neural networks trained on the same structural dataset are likely to predict the same conformation

Structure prediction with AlphaFold improved when starting from an initial guess, rather than default initialization

The AlphaFold and ESMFold databases can be mined for alternate conformations

pTM is less lenient than pLDDT toward structural novelty when predicting alternate conformations

Biomolecular diffusion models cannot reproduce the equilibrium dynamics of the simulations they are trained on

Conformational sampling by AlphaFold3-generation methods can be achieved by scaling the pair representation

Deep learning methods produce different CDRH3 conformers

Guidance potentials can be added to diffusion-based structure prediction for enhanced sampling of protein conformations

Masking ESMFold can sometimes sample alternate conformations